Advanced Concrete Technology

Advanced Concrete Technology

Constituent Materials ISBN 0 7506 5103 2Concrete Properties ISBN 0 7506 5104 0Processes ISBN 0 7506 5105 9Testing and Quality ISBN 0 7506 5106 7

Advanced Concrete TechnologyConstituent Materials

Edited byJohn NewmanDepartment of Civil EngineeringImperial CollegeLondon

Ban Seng ChooSchool of the Built EnvironmentNapier UniversityEdinburgh

AMSTERDAM BOSTON HEIDELBERG LONDON NEW YORK OXFORD

PARIS SAN DIEGO SAN FRANCISCO SINGAPORE SYDNEY TOKYO

Butterworth-HeinemannAn imprint of ElsevierLinacre House, Jordan Hill, Oxford OX2 8DP200 Wheeler Road, Burlington MA 01803

First published 2003

Copyright © 2003, Elsevier Ltd. All rights reserved

No part of this publication may be reproduced in any materialform (including photocopying or storing in any medium byelectronic means and whether or not transiently or incidentallyto some other use of this publication) without the writtenpermission of the copyright holder except in accordance with theprovisions of the Copyright, Designs and Patents Act 1988 or underthe terms of a licence issued by the Copyright Licensing Agency Ltd,90 Tottenham Court Road, London, England W1T 4LP. Applications forthe copyright holder’s written permission to reproduce any partof this publication should be addressed to the publisher

Permissions may be sought directly from Elsevier’s Science and Technology RightsDepartment in Oxford, UK: phone: (+44) (0) 1865 843830; fax: (+44) (0) 1865853333; e-mail: permissions@elsevier.co.uk. You may also complete your requeston-line via the Elsevier homepage (http://www.elsevier.com), by selecting‘Customer Support’ and then ‘Obtaining Permissions’

British Library Cataloguing in Publication DataA catalogue record for this book is available from the British Library

Library of Congress Cataloguing in Publication DataA catalogue record for this book is available from the Library of Congress

ISBN 0 7506 5103 2

For information on all Butterworth-Heinemannpublications visit our website at www.bh.com

Typeset by Replika Press Pvt Ltd, IndiaPrinted and bound in Great Britain

Contents

Preface xiiiList of contributors xv

Part 1 Cements

1 Cements 1/3Graeme Moir

1.1 Introduction 1/31.2 History of Portland cement manufacture 1/31.3 Chemistry of clinker manufacture 1/5

1.3.1 Raw materials 1/51.3.2 The modern rotary kiln 1/61.3.3 Clinkering reactions and the materials present in Portland

cement clinker 1/81.3.4 The control ratios 1/101.3.5 Calculation of clinker compound composition 1/121.3.6 Influence of minor constituents 1/12

1.4 Cement grinding 1/141.5 Portland cement hydration 1/15

1.5.1 Introduction 1/151.5.2 Hydration of silicates 1/151.5.3 Hydration of C3A and C4AF 1/161.5.4 Hydration of Portland cement 1/171.5.5 Optimization of level of rapidly soluble calcium sulfate 1/201.5.6 Techniques used to study hydration 1/211.5.7 Constitution of hydrated cement paste 1/22

1.6 Portland cement types 1/221.6.1 Standards 1/221.6.2 Main cement types 1/231.6.3 The European Standard for Common Cements (EN 197-1) 1/271.6.4 Other European cement standards 1/321.6.5 Other national standards 1/32

1.7 Cement production quality control 1/331.8 Influence of cement quality control parameters on properties 1/36

1.8.1 Key parameters 1/361.8.2 Water demand (workability) 1/361.8.3 Setting time 1/371.8.4 Strength development 1/37

1.9 Relationship between laboratory mortar results and field concrete 1/421.10 Applications for different cement types 1/421.11 Health and safety aspects of cement use 1/43References 1/45

2 Calcium aluminate cements 2/1Karen Scrivener

2.1 Introduction 2/12.1.1 Terminology 2/3

2.2 Chemistry and mineralogy of CACs 2/32.2.1 Basics of hydration and conversion 2/62.2.2 Impact of conversion 2/9

2.3 Properties of fresh CAC concrete – setting, workability, heat evolution 2/122.3.1 Setting 2/122.3.2 Workability 2/122.3.3 Rate of reaction and heat evolution 2/132.3.4 Shrinkage 2/14

2.4 Strength development 2/152.4.1 Importance of water-to-cement ratio and control of concrete quality 2/162.4.2 Other factors affecting conversion and strength development 2/18

2.5 Other engineering properties 2/192.6 Supplementary cementing materials 2/192.7 Durability/resistance to degradation 2/19

2.7.1 Reinforcement corrosion 2/202.7.2 Sulfate attack 2/202.7.3 Freeze–thaw damage 2/212.7.4 Alkaline hydrolysis 2/21

2.8 Structural collapses associated with CAC concrete 2/212.9 Modern uses of CAC concrete 2/23

2.9.1 Resistance to acids/sewage networks 2/232.9.2 Resistance to abrasion and impact 2/242.9.3 Rapid strength development 2/252.9.4 Thermal resistance 2/25

2.10 Use of CACs in mixed binder systems 2/262.10.1 Stability of ettringite-containing systems 2/282.10.2 Dimensional stability of ettringite-containing systems and

shrinkage compensations 2/28

vi Contents

2.11 Summary 2/29References 2/30

Part 2 Cementitious Additions

3 Cementitious additions 3/3Robert Lewis, Lindon Sear, Peter Wainwright and Ray Ryle

3.1 The pozzolanic reaction and concrete 3/33.2 Fly ash as a cementitious addition to concrete 3/3

3.2.1 The particle shape and density 3/63.2.2 Variability of fly ash 3/73.2.3 Fineness, water demand and pozzolanic activity 3/83.2.4 Heat of hydration 3/103.2.5 Setting time and formwork striking times 3/103.2.6 Elastic modulus 3/113.2.7 Creep 3/123.2.8 Tensile strain capacity 3/123.2.9 Coefficient of thermal expansion 3/123.2.10 Curing 3/123.2.11 Concrete durability and fly ash 3/133.2.12 Alkali–silica reaction 3/133.2.13 Carbonation of concrete 3/143.2.14 Sea water and chloride attack on reinforcing 3/143.2.15 Freeze–thaw damage 3/153.2.16 Sulfate attack 3/163.2.17 Thaumasite form of sulfate attack 3/163.2.18 Resistance to acids 3/17

3.3 Fly ash in special concretes 3/173.3.1 Roller-compacted concrete 3/173.3.2 High fly ash content in concrete 3/173.3.3 Sprayed concrete with pfa 3/17

3.4 Natural pozzolanas 3/183.5 The use of ggbs in concrete 3/18

3.5.1 Introduction 3/183.5.2 Production of ggbs 3/183.5.3 Chemical composition of ggbs 3/193.5.4 Chemical reactions 3/193.5.5 Blended cement production 3/213.5.6 Properties of concrete made with slag cements 3/223.5.7 Concluding remarks 3/34

3.6 Silica fume for concrete 3/343.6.1 The material 3/343.6.2 Inclusion in concrete 3/383.6.3 Hardened concrete: mechanical properties 3/433.6.4 Hardened concrete: durability-related properties 3/453.6.5 Concluding summary 3/48

3.7 Metakaolin 3/493.7.1 Occurrence and extraction 3/49

Contents vii

3.7.2 Metakaolin production 3/493.7.3 Physical properties 3/493.7.4 Reaction mechanisms 3/493.7.5 National standards 3/513.7.6 The effect of metakaolin on the properties of concrete 3/513.7.7 The durability of metakaolin concrete 3/543.7.8 Compatibility with blended cements 3/573.7.9 Efflorescence 3/573.7.10 Summary 3/57

3.8 Limestone 3/583.8.1 Limestone filler 3/58

References 3/59

Part 3 Admixtures

4 Admixtures for concrete, mortar and grout 4/3John Dransfield

4.1 Introduction 4/34.1.1 Definition and descripton 4/34.1.2 Brief history of admixture use 4/44.1.3 Admixture standards and types 4/44.1.4 Admixture mechanism of action 4/54.1.5 Rheology and admixtures 4/64.1.6 Cement and concrete chemistry in relation to admixtures 4/7

4.2 Dispersing admixtures 4/94.2.1 Normal plasticizers 4/94.2.2 Superplasticizers 4/104.2.3 Chemical structure of dispersing admixtures 4/104.2.4 Dispersing mechanism 4/134.2.5 Normal plasticizer performance and applications 4/154.2.6 Superplasticizer performance and applications 4/16

4.3 Retarding and retarding plasticizers/superplasticizing admixtures 4/174.3.1 Mechanism of retardation 4/174.3.2 Workability retention 4/184.3.3 Set retardation 4/194.3.4 Retarder performance and applications 4/19

4.4 Accelerating admixtures 4/204.4.1 Mechanism of acceleration 4/214.4.2 Accelerator performance and applications 4/21

4.5 Air-entraining admixtures 4/214.5.1 Factors affecting air entrainment 4/224.5.2 Freeze–thaw resistance 4/234.5.3 Air entrainment to reduce bleed 4/244.5.4 Compaction and adhesion of low-workability mixes 4/244.5.5 Bedding mortars and renders 4/25

4.6 Water resisting (waterproofing) 4/254.6.1 Mechanism 4/264.6.2 Admixture selection 4/26

viii Contents

4.7 Corrosion-inhibiting admixtures 4/264.7.1 Mechanism 4/274.7.2 Use of corrosion inhibitors 4/27

4.8 Shrinkage-reducing admixtures 4/284.8.1 Mechanism 4/284.8.2 Use of shrinkage-reducing admixtures 4/29

4.9 Anti-washout/underwater admixtures 4/294.9.1 Mechanism 4/294.9.2 Use 4/30

4.10 Pumping aids 4/304.11 Sprayed concrete admixtures 4/314.12 Foamed concrete and CLSM 4/314.13 Other concrete admixtures 4/32

4.13.1 Polymer dispersions 4/324.13.2 Pre-cast semi-dry admixtures 4/324.13.3 Truck washwater admixtures 4/32

4.14 Mortar admixtures 4/334.15 Grout admixtures 4/344.16 Admixture supply 4/34

4.16.1 Suppliers 4/344.16.2 Storage 4/354.16.3 Dispensers 4/354.16.4 Time of admixture addition 4/35

4.17 Health and safety 4/36Further reading 4/36

Part 4 Aggregates

5 Geology, aggregates and classification 5/3Alan Poole and Ian Sims

5.1 Introduction 5/35.2 Fundamentals 5/45.3 Geological classification of rocks 5/65.4 Sources and types of aggregates 5/12

5.4.1 Temperate fluvial environments 5/165.4.2 Glacial and periglacial regions 5/175.4.3 Hot desert regions 5/205.4.4 Tropical hot wet environments 5/21

5.5 Classification of aggregates 5/215.6 Aggregate quarry assessment 5/235.7 Deleterious materials in aggregates 5/25

5.7.1 Interference with the setting of the concrete 5/265.7.2 Modification to the strength and durability characteristics

of a concrete 5/295.7.3 Unsound aggregate particles in concrete 5/29

References 5/34

Contents ix

6 Aggregate prospecting and processing 6/1Mark Murrin-Earp

6.1 Aims and objectives 6/16.2 Introduction 6/16.3 Extraction and processing of sand and gravel 6/26.4 Processing 6/56.5 Extraction and processing of limestone 6/106.6 Summary 6/11Further reading 6/11

7 Lightweight aggregate manufacture 7/1P.L. Owens and J.B. Newman

7.1 Introduction, definitions and limitations 7/17.2 Lightweight aggregates suitable for use in structural concrete 7/27.3 Brief history of lightweight aggregate production 7/37.4 Manufacturing considerations for structural grades of lightweight

aggregate 7/47.4.1 Investment 7/47.4.2 Resources materials 7/47.4.3 Processes of lightweight aggregate manufacture 7/57.4.4 Production techniques 7/5

7.5 Production methods used for various lightweight aggregates 7/67.5.1 Foamed slag aggregate 7/67.5.2 Pelletized expanded blastfurnace slag aggregate 7/67.5.3 Sintered pulverized fuel ash (PFA) aggregate 7/77.5.4 Expanded clay aggregate 7/87.5.5 Expanded shale and slate aggregate 7/9

7.6 The future 7/107.7 Conclusions 7/11References 7/12

8 The effects of natural aggregates on theproperties of concrete 8/1John Lay

8.1 Aims and objectives 8/18.2 Brief history 8/18.3 Introduction 8/28.4 Classification 8/28.5 Sampling 8/28.6 Grading 8/38.7 Maximum size of aggregate 8/68.8 Aggregate shape and surface texture 8/68.9 Aggregate strength 8/88.10 Aggregate density 8/88.11 Drying shrinkage 8/98.12 Soundness 8/10

x Contents

8.13 Thermal properties 8/118.14 Fines content 8/118.15 Impurities 8/128.16 Summary 8/13References 8/14

Further reading 8/15

Index I/1

Contents xi

This Page Intentionally Left Blank

Preface

The book is based on the syllabus and learning objectives devised by the Institute ofConcrete Technology for the Advanced Concrete Technology (ACT) course. The firstACT course was held in 1968 at the Fulmer Grange Training Centre of the Cement andConcrete Association (now the British Cement Association). Following a re-organizationof the BCA the course was presented at Imperial College London from 1982 to 1986 andat Nottingham University from 1996 to 2002. With advances in computer-basedcommunications technology the traditional residential course has now been replaced inthe UK by a web-based distance learning version to focus more on self-learning ratherthan teaching and to allow better access for participants outside the UK. This book, aswell as being a reference document in its own right, provides the core material for thenew ACT course and is divided into four volumes covering the following general areas:

• constituent materials• properties and performance of concrete• types of concrete and the associated processes, plant and techniques for its use in

construction• testing and quality control processes.

The aim is to provide readers with an in-depth knowledge of a wide variety of topicswithin the field of concrete technology at an advanced level. To this end, the chapters arewritten by acknowledged specialists in their fields.

The book has taken a relatively long time to assemble in view of the many authors sothe contents are a snapshot of the world of concrete within this timescale. It is hoped thatthe book will be revised at regular intervals to reflect changes in materials, techniquesand standards.

John NewmanBan Seng Choo

This Page Intentionally Left Blank

List of contributors

John Dransfield38A Tilehouse Green Lane, Knowle, West Midlands B93 9EY, UK

John LayRMC Aggregates (UK) Ltd, RMC House, Church Lane, Bromsgrove, WorcestershireB61 8RA, UK

Robert LewisElkem Materials Ltd, Elkem House, 4a Corporation Street, High WycombeHP13 6TQ, UK

Graeme MoirEast View, Crutches Lane, Higham, Near Rochester, Kent ME2 3UH, UK

Mark Murrin-EarpRMC Aggregates (UK) Ltd, RMC House, Church Lane, Bromsgrove, WorcestershireB61 8RA, UK

John NewmanDepartment of Civil Engineering, Imperial College, London SW7 2BU, UK

Phil OwensRosebank, Donkey Lane, Tring, Hertfordshire HP23 4DY, UK

Alan PooleParks House, 1D Norham Gardens, Oxford OX2 6PS, UK

Ray Ryle3 Captains Gorse, Upper Basildon, Reading RG8 8SZ, UK

Karen ScrivenerEPFL, Swiss Federal Institute of Technology, Ecublens, Lausanne CH 1015,Switzerland

Lindon SearUK Quality ASL Association, Regent House, Bath Avenue, Wolverhampton, WestMidlands WV1 4EG, UK

Ian SimsSTATS Consultancy, Porterswood House, Porterswood, St Albans AL3 6PQ, UK

Peter WainwrightSchool of Civil Engineering, The University of Leeds, Leeds L52 9TT, UK

xvi List of contributors

PART 1

Cements

This Page Intentionally Left Blank

CementsGraeme Moir

1.1 Introduction

The aims and objectives of this chapter are to:

• describe the nature of Portland (calcium silicate-based) cements• outline the manufacturing process and the quality control procedures employed• review the cement hydration processes and the development of hydrated structures• outline the influence of differences in cement chemistry and compound composition

on the setting and strength development of concrete• review cement types (including composite and masonry cements) and the nature of

their constituents• review the standards with which cements must comply and the applications for different

cement types• describe in outline methods used for chemical analysis and to study the hydration of

cement• briefly outline some health and safety aspects related to cement use

1.2 History of Portland cement manufacture

Portland cement is essentially a calcium silicate cement, which is produced by firing topartial fusion, at a temperature of approximately 1500°C, a well-homogenized and finelyground mixture of limestone or chalk (calcium carbonate) and an appropriate quantityof clay or shale. The composition is commonly fine tuned by the addition of sand and/oriron oxide.

1

1/4 Cements

The first calcium silicate cements were produced by the Greeks and Romans, whodiscovered that volcanic ash, if finely ground and mixed with lime and water, produceda hardened mortar, which was resistant to weathering. The reaction is known as thepozzolanic reaction and it is the basis of the contribution made to strength and concreteperformance by materials such as fly ash, microsilica and metakaolin in modern concrete.

In the mid eighteenth century John Smeaton discovered that certain impure limes(these contained appropriate levels of silica and alumina) had hydraulic properties. Thatis, they contained reactive silicates and aluminates, which could react with water to yielddurable hydrates, which resisted the action of water. Smeaton used this material in themortar used to construct the Eddystone Lighthouse in 1759.

The term ‘Portland cement’ was first applied by Joseph Aspdin in his British PatentNo. 5022 (1824), which describes a process for making artificial stone by mixing limewith clay in the form of a slurry and calcining (heating to drive off carbon dioxide andwater) the dried lumps of material in a shaft kiln. The calcined material (clinker) wasground to produce cement. The term ‘Portland’ was used because of the similarity of thehardened product to that of Portland stone from Dorset and also because this stone had anexcellent reputation for performance.

Joseph Aspdin was not the first to produce a calcium silicate cement but his patentgave him the priority for the use of the term ‘Portland cement’. Other workers were activeat the same time or earlier, most notably Louis Vicat in France. Blezard (1998) gives acomprehensive review of the history of the development of calcareous (lime-based) cements.

The cements produced in the first half of the nineteenth century did not have the samecompound composition as modern Portland cements as the temperature achieved was nothigh enough for the main constituent mineral of modern cements, tricalcium silicate(C3S), to be formed. The only silicate present was the less reactive dicalcium silicate(C2S). The sequences of reactions, which take place during clinker production, are discussedin section 1.3.3.

The main technical innovations in cement manufacture which have taken place overapproximately the last 150 years are summarized in Figure 1.1.

It was the introduction of the rotary kiln at the end of the nineteenth century thatenabled a homogeneous product to be manufactured, which had experienced a consistentlyhigh enough temperature to ensure C3S formation. During the twentieth century thenature of the product changed relatively little in terms of its overall chemistry and mineralcomposition but there have been considerable advances in production technology resultingin improved energy efficiency, improved quality control, reduced environmental impactand lower labour intensity.

It should be noted that the introduction of rotary kiln technology in the early twentiethcentury coincided with the publication of cement standards in the UK and the USA. Bothstandards required the strength of a briquette of cement paste to reach minimum valuesat 7 and 28 days.

The control of clinker composition has advanced from the volume proportions arrivedat by trial and error in the late nineteenth century to precise control using rapid X-rayfluorescence techniques. The continuous improvements in manufacturing methods andquality control combined with market competitive pressures have resulted in a fourfoldincrease in the 28-day strength given by a typical European Portland cement at 28 dayssince the late nineteenth century (Blezard, 1998). In Europe, this strength escalation haseffectively been controlled by the introduction of cement standards with upper as well as

Cements 1/5

lower strength limits. The European Standard for Common Cements (EN 197-1) is outlinedin section 1.6.3.

Figure 1.1 Landmarks in Portland cement production.

2000–

1980

1960

1940

1920

1900

1880

1860

1840

1820

Multistage combustion – emission control

New horizontal cement mill technology

High-pressure roll press for cement pregrinding

Automatic kiln control using expert systems

High-efficiency separator introduced for cement grinding

Precalciner process developed

X-ray fluorescence (XRF) rapid chemical analysis

Suspension preheater process introduced

Lepol (nodule) process introduced

Introduction of pneumatic blending silos for raw meal

First electrostatic precipitator installed in cement works

Paper sacks introduced for cement

British (BS 12) and ASTM (C9) standards published

Rotary kilns introduced

Tube grinding mills for cement

Method for carbonate of lime developed

J. Grant introduced tensile strength test for cement

W. Aspdin bottle kiln plant at Northfleet

Patent for Portland Cement granted to J. Aspdin

1.3 Chemistry of clinker manufacture

1.3.1 Raw materials

Cement making is essentially a chemical process industry and has much in common withthe manufacture of so-called heavy chemicals such as sodium hydroxide and calciumchloride. Close control of the chemistry of the product is essential if cement with consistentproperties is to be produced. This control applies not only to the principal oxides whichare present but also to impurities, which can have a marked influence on both themanufacturing process and cement properties.

As illustrated in Figure 1.2, a chemical analysis of Portland cement clinker shows it toconsist mainly of four oxides: CaO (lime), SiO2 (silica), Al2O3 (alumina) and Fe2O3 (ironoxide). In order to simplify the description of chemical composition, a form of shorthandis used by cement chemists in which the four oxides are referred to respectively as C, S,A and F.

Expressing the chemical analysis in the form of oxides, rather than the individualelements of silicon (Si), calcium (Ca) etc., has the advantage that the analysis total should

1/6 Cements

come close to 100, and this provides a useful check for errors. Oxidizing conditions aremaintained during the burning process and this ensures that the metallic elements presentare effectively present as oxides although combined in the clinker as minerals.

The source of lime for cement making is usually limestone or chalk. As typically 80%of the raw mix consists of limestone, it is referred to as the primary raw material. Thesecondary raw material, which provides the necessary silica, alumina and iron oxide, isnormally shale or clay. Small quantities of sand or iron oxide may be added to adjust thelevels of silica and iron oxide in the mix. When proportioning the raw materials, anallowance must be made for ash incorporated into the clinker from the fuel that fires thekiln. Most cement plants worldwide use finely ground (pulverized) coal as the primaryfuel. Increasingly, by-product fuels such as the residue from oil refining (petroleum coke)and vehicle tyres are being used to partially replace some of the coal.

Typical contents of the four principal oxides in a simplified cement making operationutilizing only two raw materials are given in Figure 1.3.

Note that the ratio of CaO to the other oxides is lower in the clinker than in the rawmix. This is a result of the incorporation of shale from the coal ash. The levels of theoxides are also increased as a result of decarbonation (removal of CO2).

1.3.2 The modern rotary kiln

The rotary kilns used in the first half of the twentieth century were wet process kilnswhich were fed with raw mix in the form of a slurry. Moisture contents were typically40% by mass and although the wet process enabled the raw mix to be homogenizedeasily, it carried a very heavy fuel penalty as the water present had to be driven off in thekiln.

In the second half of the twentieth century significant advances were made which haveculminated in the development of the precalciner dry process kiln. In this type of kiln, the

Figure 1.2 Typical chemical composition of Portland cement clinker.

MainCement

constituentschemist’sshorthand %

SiO2 S 21.1

Al2O3 A 5.6

Fe2O3 F 3.0

CaO C 65.5

95.2

C

FAS

Minorconstituents

Minorconstituents

%Mn2O3 0.05P2O5 0.15TiO2 0.30MgO 1.50SO3(S) 1.20Loss on ignition 0.50K2O 0.72Na2O 0.18Fluorine 0.04Chloride 0.02Trace elements 0.01

4.67

Cements 1/7

Figure 1.3 Typical chemical analyses of materials.

Limestone(95% CaCO)3

Shale

c. 80 parts c. 20 parts

%S 52.8

A 14.2

F 8.7

C 1.0

%S 3.3

A 0.7

F 0.2

C 53.5

Mill

Rawfeed

CO2 gas

Kiln

Coal

%S 51.7

A 26.4

F 9.5

C 1.6

%S 13.2

A 3.4

F 1.9

C 43.0c. 10–15% ash

Clinker

%S 20.9

A 5.6

F 3.0

C 65.7

Figure 1.4 The modern precalciner kiln.

Raw feed

Cyclones

Precalciner vessel

Back-end fuel

Kiln fuel

CoolerClinker

Rotary kiln

Fan

1/8 Cements

energy-consuming stage of decarbonating the limestone present in the raw mix is completedbefore the feed enters the rotary kiln. The precalcination of the feed brings many advantages,the most important of which is high kiln output from a relatively short and small-diameterrotary kiln. Almost all new kilns installed since 1980 have been of this type. Figure 1.4illustrates the main features of a precalciner kiln.

The raw materials are ground to a fineness, which will enable satisfactory combinationto be achieved under normal operating conditions. The required fineness depends on thenature of the raw materials but is typically in the range 10–30% retained on a 90 micronsieve. The homogenized raw meal is introduced into the top of the preheater tower andpasses downwards through a series of cyclones to the precalciner vessel. The raw meal issuspended in the gas stream and heat exchange is rapid. In the precalciner vessel the mealis flash heated to ~900°C and although the material residence time in the vessel is onlya few seconds, approximately 90% of the limestone in the meal is decarbonated beforeentering the rotary kiln. In the rotary kiln the feed is heated to ~1500°C and as a result ofthe tumbling action and the partial melting it is converted into the granular materialknown as clinker. Material residence time in the rotary kiln of a precalciner process istypically 30 minutes. The clinker exits the rotary kiln at ~1200°C and is cooled to ~60°Cin the cooler before going to storage and then being ground with gypsum (calciumsulfate) to produce cement. The air which cools the clinker is used as preheated combustionair thus improving the thermal efficiency of the process. As will be discussed in section1.5, the calcium sulfate is added to control the initial hydration reactions of the cementand prevent rapid, or flash, setting.

If coal is the sole fuel in use then a modern kiln will consume approximately 12 tonnesof coal for every 100 tonnes of clinker produced. Approximately 60% of the fuel inputwill be burned in the precalciner vessel. The high fuel loading in the static precalcinervessel reduces the size of rotary kiln required for a given output and also reduces theconsumption of refractories. A wider range of fuel types (for example, tyre chips) can beburnt in the precalciner vessel than is possible in the rotary kiln.

Although kilns with daily clinker outputs of ~9000 tonnes are in production in Asiamost modern precalciner kilns in operation in Europe have a production capability ofbetween 3000 and 5000 tonnes per day.

1.3.3 Clinkering reactions and the minerals present inPortland cement clinker

Portland cement clinker contains four principal chemical compounds, which are normallyreferred to as the clinker minerals. The composition of the minerals and their normalrange of levels in current UK and European Portland cement clinkers are summarized inTable 1.1.

It is the two calcium silicate minerals, C3S and C2S, which are largely responsible forthe strength development and the long-term structural and durability properties of Portlandcement. However, the reaction between CaO (lime from limestone) and SiO2 (silica fromsand) is very difficult to achieve, even at high firing temperatures. Chemical combinationis greatly facilitated if small quantities of alumina and iron oxide are present (typically5% Al2O3 and 3% Fe2O3), as these help to form a molten flux through which the lime andsilica are able to partially dissolve, and then react to yield C3S and C2S. The sequence ofreactions, which take place in the kiln, is illustrated in Figure 1.5.

Cements 1/9

Table 1.1 Ranges of principal minerals in European clinkers

Shorthand Chemical Mineral Typical level Typical rangenomenclature formula name (mass %) (mass %)

C3S 3CaO·SiO2 Alite 57 45–65or Ca3SiO5

C2S 2CaO·SiO2 Belite 16 10–30or Ca2SiO4

C3A 3CaO·Al2O3 Aluminate 9 5–12or Ca3Al2O6

C4AF 4CaO·Al2O3·Fe2O3 Ferrite 10 6–12or Ca4Al2Fe2O10

Figure 1.5 Sequence of reactions taking place during the formation of Portland cement clinker.(Source: Reproduced by courtesy of KHD Humbolt Wedag AG.)

CO2

Free lime

Alite Clin

ker

Belite

Liquid

LiquidC12A7

Cr

High quartzLow quartz

Clay minerals

Fe2O3 H2OC2(A,F) C4AF

0 200 400 600 800 1000 1200 1400

Temperature (°C)

Por

tions

by

wei

ght

Raw

mea

l CaCO3

The reaction requiring the greatest energy input is the decarbonation of CaCO3, whichtakes place mainly in the temperature range 700–1000°C. For a typical mix containing80% limestone the energy input to decarbonate the CaCO3 is approximately 400 kCal/kgof clinker, which is approximately half of the total energy requirement of a modern dryprocess kiln.

When decarbonation is complete at about 1100°C, the feed temperature rises morerapidly. Lime reacts with silica to form belite (C2S) but the level of unreacted limeremains high until a temperature of ~1250°C is reached. This is the lower limit ofthermodynamic stability of alite (C3S). At ~1300°C partial melting occurs, the liquidphase (or flux) being provided by the alumina and iron oxide present. The level of

C3A

1/10 Cements

unreacted lime reduces as C2S is converted to C3S. The process will be operated to ensurethat the level of unreacted lime (free lime) is below 3%.

Normally, C3S formation is effectively complete at a material temperature of about1450°C, and the level of uncombined lime reduces only slowly with further residencetime. The ease with which the clinker can be combined is strongly influenced by themineralogy of the raw materials and, in particular, the level of coarse silica (quartz)present. The higher the level of coarse silica in the raw materials, the finer the raw mixwill have to be ground to ensure satisfactory combination at acceptable kiln temperatures.

Coarse silica is also associated with the occurrence of clusters of relatively large belitecrystals around the sites of the silica particles. Figures 1.6(a) and 1.6(b) are photomicrographsof a ‘normal’ clinker containing well-distributed alite and belite and clinker producedfrom a raw meal containing relatively coarse silica.

Figure 1.6 Reflected light photomicrographs of Portland cement clinker.

Alite Belite Flux (interstitial)phase

(a)

Good quality clinker, smalluniformly sized C3S and C2S

C2S cluster caused bycoarse silica in raw mix

Belite Alite(b)

The belite present in the clusters is less reactive than small well-distributed belite andthis has an adverse influence on cement strengths.

As the clinker passes under the flame it starts to cool and the molten C3A and C4AF,which constitute the flux phase, crystallize. This crystallization is normally completeby the time the clinker exits the rotary kiln and enters the cooler at a temperature of~1200°C. Slow cooling should be avoided as this can result in an increase in the belitecontent at the expense of alite and also the formation of relatively large C3A crystalswhich can result in unsatisfactory concrete rheology (water demand and stiffening).

1.3.4 The control ratios

The control of clinker composition, and optimization of plant performance is greatlyassisted by the use of three ratios:

Lime saturation factor LSF = 2.8 + 1.2 + 0.65

100%CS A F

×

Silica ratio SR = + S

A F

Cements 1/11

Alumina ratio AR = AF

The most critical control ratio is the lime saturation factor, which is determined by theratio of lime, to silica, alumina and iron oxide, and governs the relative proportions ofC3S and C2S. The formula for LSF has been derived from high-temperature phase equilibriastudies. When the LSF is above 100% there is an excess of lime, which cannot becombined no matter how long the clinker is fired, and this remains as free lime in theclinker. As a low level of uncombined lime must be achieved (~3% maximum and preferablybelow 2%), clinker LSFs normally lie in the range 95–98%. Figure 1.7 illustrates theinfluence of LSF on the content of C3S and C2S, and the firing temperature required tomaintain clinker free lime below 2% (normally referred to as the combinability temperature).The contents of C3S and C2S have been calculated by the so-called Bogue method. Thisprocedure is discussed further in section 1.3.5. Normally, if the LSF is increased at aparticular cement plant, the raw mix must be ground finer, i.e. the percentage of particlescoarser than 90 microns is reduced.

Figure 1.7 Influence of clinker LSF on compound composition and ease of combination.

Proportions of clinker minerals

C3S

SR 2.5

AR 2.5

C2S

C3AC4AF

LSF %

80 85 90 95 100 105

Wt. %70

60

50

40

30

20

10

0

T°C1550

1500

1450

1400

1350

130090 92 94 96 98 100

LSF%

10% plus90 microns

15% plus90 microns

Burning temperature required toreduce clinker-free lime below 2%

In order to ensure optimum kiln performance and uniform cement quality it is essentialthat the LSF of the raw mix is maintained within a narrow band, ideally ±2% or, moreprecisely, with a standard deviation of better than 1%, determined on hourly samples. Asa change in LSF of 1% (at constant free lime) corresponds to a change in C3S of ~2% theC3S variability range is approximately double that of the LSF range.

The silica ratio, SR, is the ratio of silica to alumina and iron oxide. For a given LSF,the higher the silica ratio, the more C3S and the less C3A and C4AF will be produced. Ofgreater significance, with regard to clinker manufacture, is that the higher the SR, the lessmolten liquid, or flux, is formed. This makes clinker combination more difficult unlessthe LSF is reduced to compensate. The flux phase facilitates the coalescence of theclinker into nodules and also the formation of a protective coating on the refractory kilnlining. Both are more difficult to achieve as the SR increases.

The alumina ratio, AR, is normally the third ratio to be considered. An AR of ~1.4 isnormally optimum for clinker burning, as at this value the quantity of liquid phase formedwhen partial melting first occurs at ~1300oC is maximized. High AR cements will have

1/12 Cements

a high C3A content, and this can be disadvantageous in certain cement applications, forexample where it is desired to minimize the concrete temperature rise. Fortunately, theAR ratio is relatively easy to control by means of a small addition of iron oxide tothe mix.

As a result of market requirements for cements from different sources to have similarproperties and also to optimize clinker production there has been a trend to converge ona ‘standard’ clinker chemistry of

LSF 95–97%SR 2.4–2.6AR 1.5–1.8

At most plants the achievement of this ideal chemistry will require the use of correctivematerials such as sand and iron oxide.

1.3.5 Calculation of clinker compound composition

The levels of the four clinker minerals can be estimated using a method of calculationfirst proposed by Bogue in 1929 (see Bogue, 1955). The method involves the followingassumptions:

• all the Fe2O3 is combined as C4AF• the remaining Al2O3 (i.e. after deducting that combined in C4AF) is combined as C3A

The CaO combined in the calculated levels of C3A and C4AF and any free lime arededucted from the total CaO and the level of SiO2 determines the proportions of C3S andC2S. The procedure can be expressed mathematically (in mass%) as follows:

C3S = 4.071(total CaO – free lime) – 7.600SiO2 – 6.718Al2O3 – 1.430Fe2O3

C2S = 2.867SiO2 – 0.7544C3S

C3A = 2.65Al2O3 – 1.692Fe2O3

C4AF = 3.043Fe2O3

The calculated figures may not agree exactly with the proportions of the clinker mineralsdetermined by quantitative X-ray diffraction or by microscopic point counting. However,they give a good guide to cement properties in terms of strength development, heat ofhydration and sulfate resistance.

When calculating the compound composition of cements (rather than clinkers) thenormal convention is to assume that all the SO3 present is combined with Ca (i.e. ispresent as calcium sulfate). The total CaO is thus reduced by the free lime level andby 0.7 × SO3. Examples of the calculation for cements are given in Table 1.7 later inthis chapter.

1.3.6 Influence of minor constituents

As illustrated in Figure 1.2, approximately 95% of clinker consists of the oxides of CaO,SiO2, Al2O3 and Fe2O3 (but present in combined form as the clinker minerals) and the

Cements 1/13

remainder consists of the so-called minor constituents. The influence of minor constituentson cement manufacture and cement properties has been reviewed (Moir and Glasser,1992; Bhatty, 1995).

Table 1.2 indicates the typical UK levels of the most commonly encountered minorconstituents and summarizes their impact on the cement manufacturing process. Theinputs of alkali metal oxides (Na2O and K2O), SO3 and chloride have to be closelycontrolled because they are volatilized in the kiln and can cause severe operational problemsassociated with their condensation and the formation of build-ups in the kiln ‘back end’and preheater.

Table 1.2 Influence of most commonly encountered minor constituents on manufacturing process

Minor constituent Typical range of levels in UK Influence on processclinkers

Na2O 0.07–0.22 Alkali sulfates are volatilized in thekiln and condense in lower-temperature

K2O 0.52–1.0 regions resulting in build-ups andSO3 0.5–1.5 blockages

Fluorine 0.01–0.20 Greatly assists combination by virtueof mineralizing action

Chloride 0.005–0.05 Alkali chlorides are highly volatile andcause build-ups and blockages

MgO 0.8–2.5 Slight fluxing action

Trace metals 5–100ppm Slight – but some (e.g. thallium) haveto be minimized to limit emissions tothe environment

The alkali metals Na2O and K2O have a very strong affinity for SO3 and a liquid phasecontaining Na+, K+, Ca2+ and SO4

2– ions is formed which is immiscible with the mainclinker liquid (molten C3A and C4AF). On cooling this crystallizes to yield alkali sulfatessuch as K2SO4, aphthitalite (3K2SO4·Na2SO4) and calcium langbeinite (K2SO4· 2CaSO4).The crystallization products depend on the relative levels of the two alkali oxides and thelevel of SO3. If there is insufficient SO3 to combine with the alkali metal oxides thenthese may enter into solid solution in the aluminate and silicate phases. C2S can bestabilized at temperatures above 1250oC thus impeding the formation of C3S. A similarstabilization of C2S, requiring ‘hard burning’ to lower the free lime level to an acceptablelevel, can occur if there is a large excess of SO3 over alkalis.

A deficiency of SO3 in the clinker is associated with enhanced C3A activity anddifficulties in achieving satisfactory early age concrete rheology.

Fluorine occurs naturally in some limestone deposits, for example in the Pennines inEngland, and has a beneficial effect on clinker combination. It acts as both a flux andmineralizer, increasing the quantity of liquid formed at a given temperature and stabilizingC3S below 1250oC. The level in the clinker, however, must be controlled below ~ 0.25%in order to avoid a marked reduction in the early reactivity of cement.

Minor constituents also have to be controlled on account of their impact on cementproperties and also concrete durability. Related to this, the levels of alkalis, SO3, chlorideand MgO are also limited by national cement standards or codes of practice. Theseaspects are reviewed in section 1.6.

1/14 Cements

1.4 Cement grinding

The clinker is normally conveyed to a covered store where, provided stocks are adequate,it will cool from the cooler discharge temperature of 50–80°C to a temperature approachingambient. If clinker stocks are low then the clinker may be ground to cement without theopportunity to cool further during storage. The clinker is ground to a fine powder withapproximately 5% calcium sulfate, which is added to control the early reactions of thealuminate phase. These reactions are considered in detail in section 1.5. The Europeanstandard for common cements also permits the addition of up to 5% of a minor additionalconstituent (mac), which, in practice, is normally limestone. Macs can be helpful inoptimizing cement rheological properties.

The vast majority of cement produced throughout the world is ground in ball mills,which are rotating tubes containing a range of sizes of steel balls. A closed-circuit millinginstallation is illustrated in Figure 1.8.

Figure 1.8 Schematic diagram of closed-circuit grinding mill.

Fan

Dust

Classifier

Dustfilter

Bucketelevator

Product

Dischargehood

Mill

Rejects

Feeder

Feed

Closed-circuit mills normally have two chambers separated by a slotted diaphragm,which allows the partially ground cement to pass through but retains the grinding balls.The first chamber contains large steel balls (60–90 mm in diameter), which crush theclinker. Ball sizes in the second chamber are normally in the range 19–38 mm.

The mill operates in a closed circuit in which the mill product passes to a separatingdevice where coarse particles are rejected and returned to the mill for further grinding.The final product can thus be significantly finer than the material that exits the mill. Millswhich do not have this separating stage are known as open-circuit mills and they are lessefficient particularly at high cement finenesses (above 350 m2/kg).

The efficiency of the clinker grinding process is very low. Less than 2% of the electricalenergy consumed is used in actually fracturing the particles; the rest is converted to heat.Modern mills are equipped with internal water sprays, which cool the process by evaporation.Cement mill temperatures are typically in the range 110–130°C and at this temperaturethe hydrated form of calcium sulfate (gypsum, CaSO4·2H2O) added to control the initial

Cements 1/15

hydration reactions undergoes dehydration. This has some advantages but the level ofdehydrated calcium sulfate has to be controlled to optimize the water demand propertiesof the cement. This aspect is discussed in section 1.5.5.

The low efficiency of the grinding process has resulted in considerable effort beingdirected to find more efficient processes. Some of these developments are listed in Figure1.1. As a general rule the more efficient the grinding process, the steeper the particle sizegrading. The range of particle sizes is smaller and this can result in increased waterdemand of the cement, at least in pastes and rich concrete mixes. This is because with anarrow size distribution there are insufficient fine particles to fill the voids between thelarger particles and these voids must be filled by water. Concerns over product performancein the market and also mechanical/maintenance problems with some of the new millingtechnologies have resulted in the ball mill retaining its dominant position. One compromise,which lowers grinding power requirement without prejudicing product quality, is theinstallation of a pre-grinder, such as a high-pressure roll press, to finely crush the clinkerobviating the need for large grinding media in the first chamber of the ball mill.

1.5 Portland cement hydration

1.5.1 Introduction

The hydration of Portland cement involves the reaction of the anhydrous calcium silicateand aluminate phases with water to form hydrated phases. These solid hydrates occupymore space than the anhydrous particles and the result is a rigid interlocking mass whoseporosity is a function of the ratio of water to cement (w/c) in the original mix. Providedthe mix has sufficient plasticity to be fully compacted, the lower the w/c, the higher willbe the compressive strength of the hydrated cement paste/mortar/concrete and the higherthe resistance to penetration by potentially deleterious substances from the environment.

Cement hydration is complex and it is appropriate to consider the reactions of thesilicate phases (C3S and C2S) and the aluminate phases (C3A and C4AF) separately. Thehydration process has been comprehensively reviewed (Taylor, 1997).

1.5.2 Hydration of silicates

Both C3S and C2S react with water to produce an amorphous calcium silicate hydrateknown as C–S–H gel which is the main ‘glue’ which binds the sand and aggregateparticles together in concrete. The reactions are summarized in Table 1.3. C3S is muchmore reactive than C2S and under ‘standard’ temperature conditions of 20°C approximatelyhalf of the C3S present in a typical cement will be hydrated by 3 days and 80% by 28days. In contrast, the hydration of C2S does not normally proceed to a significant extentuntil ~14 days.

The C–S–H produced by both C3S and C2S has a typical Ca to Si ratio of approximately1.7. This is considerably lower than the 3:1 ratio in C3S and the excess Ca is precipitatedas calcium hydroxide (CH) crystals. C2S hydration also results in some CH formation.The following equations approximately summarize the hydration reactions:

1/16 Cements

Several theories have been developed to explain this dormant period. The most favouredis that the initial reaction forms a protective layer of C–S–H on the surface of the C3S andthe dormant period ends when this is destroyed or rendered more permeable by ageing ora change in structure. Reaction may also be inhibited by the time taken for nucleation ofthe C–S–H main product once water regains access to the C3S crystals.

1.5.3 Hydration of C3A and C4AF

The reactions of laboratory-prepared C3A and C4AF with water, alone or in the presenceof calcium sulfate and calcium hydroxide have been extensively studied (Odler, 1998).However, the findings should be interpreted with caution as the composition of thealuminate phases in industrial clinker differs considerably from that in synthetic preparationsand hydration in cements is strongly influenced by the much larger quantity of silicatesreacting and also by the presence of alkalis.

In the absence of soluble calcium sulfate C3A reacts rapidly to form the phases C2AH8

and C4AH19, which subsequently convert to C2AH6. This is a rapid and highly exothermicreaction.

If finely ground gypsum (CaSO4·2H2O) or hemihydrate (CaSO4·0.5H2O) is blendedwith the C3A prior to mixing with water then the initial reactions are controlled by theformation of a protective layer of ettringite on the surface of the C3A crystals. Thereaction can be summarized as:

C3A + 3C + 3 S + 32H ⇒ C3A · 3C S · 32H

where in cement chemists’ notation S represents SO3 and H represents H2O, i.e.

C3A + dissolved calcium (Ca2+) + dissolved sulfate (SO )42– + water ⇒ ettringite

The more rapid dissolution of dehydrated forms of gypsum ensures an adequate supplyof dissolved calcium and sulfate ions and will be more effective in controlling the reactionof finely divided or highly reactive forms of C3A. The role of gypsum dehydration isconsidered further in section 1.5.5.

In most commercial Portland cements there will be insufficient sulfate available to

Table 1.3 Hydration of calcium silicates

Mineral Reaction rate Products of reaction

C3S Moderate C–S–H with Ca:Si ratio ~ 1.7CH (calcium hydroxide)

C2S Slow C–S–H with Ca:Si ratio ~ 1.7Small quantity of CH

C3S + 4.3H ⇒ C1.7SH3 + 1.3CH

C2S + 3.3H ⇒ C1.7SH3 + 0.3CH

An important characteristic of C3S hydration is that after an initial burst of reactionwith water on first mixing it passes through a dormant, or induction, period where reactionappears to be suspended. This is of practical significance because it allows concrete to beplaced and compacted before setting and hardening commences.

Cements 1/17

sustain the formation of ettringite. When the available sulfate has been consumed theettringite reacts with C3A to form a phase with a lower SO3 content known as monosulfate.The reaction can be summarized as:

C3A · 3C S · 32H + 2C3A + 4H ⇒ 3(C3A · C S ·12H)

Many studies have shown that the hydration of C4AF (or more correctly the C2A–C2Fsolid solution) is analogous to that of C3A but proceeds more slowly (Taylor, 1997). Theiron enters into solid solution in the crystal structures of ettringite and monosulfatesubstituting for aluminium. In order to reflect the variable composition of ettringite andmonosulfate formed by mixtures of C3A and C4AF they are referred to respectively asAFt (alumino-ferrite trisulfate hydrate) and AFm (alumino-ferrite monosulfate hydrate)phases. The hydration reactions of C3A and C4AF are summarized in Table 1.4.

Table 1.4 Hydration of aluminates

Mineral Soluble calcium Reaction rate Products of reactionsulfate present

C3A No Very rapid with Hydrates of type C2AH8 andrelease of heat C4AH19 which subsequently convert

to C2AH6

C3A Yes Initially rapid Ettringite C3A3C S 32H whichsubsequently reacts to formmonosulfate 3(C3A· C S ·12H)

C4AF No Variable (depends on Hydrates of type C2(A,F)H8 andAl/Fe ratio) C4(A,F)Hx which subsequently

convert to C3(A,F)H6 (hydrogarnet)

C4AF Yes Variable but generally Iron substituted ettringite (AFt) andslow subsequently iron substituted

monosulfate (AFm)

1.5.4 Hydration of Portland cement

The hydration of Portland cement is rather more complex than that of the individualconstituent minerals described above. A simplified illustration of the development ofhydrate structure in cement paste is given in Figure 1.9.

When cement is first mixed with water some of the added calcium sulfate (particularlyif dehydrated forms are present, and most of the alkali sulfates present (see section 1.3.5),dissolve rapidly. If calcium langbeinite is present then it will provide both calcium andsulfate ions in solution, which are available for ettringite formation.

The supply of soluble calcium sulfate controls the C3A hydration, thus preventing aflash set. Ground clinker mixed with water without added calcium sulfate sets rapidlywith heat evolution as a result of the uncontrolled hydration of C3A. The cement thenenters a dormant period when the rate of loss of workability is relatively slow. It will bemore rapid, however, at high ambient temperatures (above 25°C).

Setting time is a function of clinker mineralogy (particularly free lime level), clinkerchemistry and fineness. The finer the cement and the higher the free lime level, theshorter the setting time in general. Cement paste setting time is arbitrarily defined as thetime when a pat of cement paste offers a certain resistance to penetration by a probe of

1/18 Cements

standard cross-section and weight (see section 1.7). Setting is largely due to the hydrationof C3S and it represents the development of hydrate structure, which eventually results incompressive strength.

The C–S–H gel which forms around the larger C3S and C2S grains is formed in situand has a rather dense and featureless appearance when viewed using an electron microscope.This material is formed initially as reaction rims on the unhydrated material but ashydration progresses the anhydrous material is progressively replaced and only the largestparticles (larger than ~30 microns) will retain an unreacted core after several years’hydration. This dense hydrate is referred to as the ‘inner product’.

The ‘outer hydration product’ is formed in what was originally water-filled space andalso space occupied by the smaller cement grains and by interstitial material (C3A andC4AF). When viewed using an electron microscope this material can be seen to containcrystals of Ca(OH)2, AFm/AFt and also C–S–H with a foil- or sheet-like morphology.The structure of the outer product is strongly influenced by the initial water-to-cementratio, which in turn determines paste porosity and consequently strength development.

Figure 1.9 Simplified illustration of hydration of cement paste.

C2S C3A

Ferrite

~ 15 minutesCement mixed with water

Calcium sulfatedissolves

Calciumsulfate

C3S

Fresh cement

Protective layerof ettringite

formed on C3A

~3 hoursCement paste starts to set

C–S–H hydrateforms on C3S

Calciumhydroxide(portlandite)precipitated

~28 days

C2S formsadditionalC–S–H hydrate

Ferrite yieldssimilar hydrateto C3A (AFm)

Cement paste hardenedUnreactedcentres of

coarse particlesPorositydepends onwater:cementratio

C3S C2S C3A C4AP C–S–H Aluminate Calcium Calciumgel hydrate hydroxide sulfate

Cements 1/19

The hydration of Portland cement involves exothermic reactions, i.e. they release heat.The progress of the reactions can be monitored using the technique of isothermal conductioncalorimetry (Killoh, 1988).

Figure 1.10 Heat of hydration of a cement paste determined by conduction calorimetry at 20°C.

500

400

300

200

100

0

Tota

l hea

t ev

olve

d (k

J/kg

)

Renewed aluminate phasehydration

Total heat evolved

1 5 10 25 50

Initial reactions ofwetting andformation ofprotective coatings

Main C3S hydration peak(formation of C–S–H and CHprecipitation)

5

4

3

2

1

0

Rat

e of

hea

t ev

olut

ion

(W/k

g)

Hydration time (hours)

The shoulder on the main hydration peak which is often seen at ~16 hours is associatedwith renewed ettringite formation which is believed to occur as a result of instability ofthe ettringite protective layer. In some cements with a low ratio of SO3 to C3A it may beassociated with the formation of monosulfate.

The heat release is advantageous in cold weather and in precast operations where thetemperature rise accelerates strength development and speeds up the production process.However, in large concrete pours the temperature rise, and in particular the temperaturedifference between the concrete core and the surface can generate stresses which result in‘thermal cracking’. Figure 1.11 illustrates the influence of concrete pour size on concretetemperature for a typical UK Portland cement. The data were obtained using the equipmentdescribed by Coole (1988).

The temperature rise experienced depends on a number of factors, which include:

• concrete placing temperature• cement content• minimum pour dimensions• type of formwork• cement type (fineness, C3S and C3A contents)

Cement heat of hydration (during the first ~ 48 hours) is highest for finely ground cementswith a high C3S content (>60%) and a high C3A content (>10%).

By 28 days a typical Portland cement cured at 20°C can be expected to be ~90%hydrated. The extent of hydration is strongly influenced by cement fineness and in particularthe proportions of coarse particles in the cement. Cement grains which are coarser than~30 microns will probably never fully hydrate. Thus, cement particle size distribution has

1/20 Cements

a strong influence on long-term compressive strength. Cement produced in an open-circuit mill with a 45 micron sieve residue of 20% may give a 28-day strength ~10%lower than that of a cement produced from the same clinker but ground in a closed-circuitmill with a 45 micron sieve residue of 3% (Moir, 1994).

Elevated temperature curing, arising from either the semi-adiabatic conditions existingin large pours or from externally applied heat, is associated with reduced ultimate strength.This is believed to be due to a combination of microcracks induced by thermal stressesbut also a less dense and ‘well-formed’ microstructure.

1.5.5 Optimization of level of rapidly soluble calcium sulfate

As described in section 1.3 cement mill temperatures normally lie in the range 100–130°C. Under these conditions the calcium sulfate dihydrate (gypsum) added to the millundergoes dehydration first to hemihydrate (CaSO4·0.5H2O) and then to soluble anhydrite(CaSO4). These dehydrated forms of gypsum are present in commercial plasters and it isthe formation of an interlocking mass of gypsum crystals which is responsible for thehardening of plaster once mixed with water.

The dehydrated forms of gypsum dissolve more rapidly than gypsum and this isbeneficial in ensuring that sufficient Ca2+ and SO 4

2– ions are available in solution tocontrol the initial reactivity of C3A by forming a protective layer of ettringite. An inadequatesupply of soluble calcium sulfate can result in a rapid loss of workability known as flashset. This is accompanied by the release of heat and is irreversible.

However, if too high a level of dehydrated gypsum is present, then crystals of gypsumcrystallize from solution and cause a plaster or false set. This is known as false setbecause if mixing continues, or is resumed, the initial level of workability is restored.

Figure 1.11 Influence of pour size (minimum dimensions) on concrete temperature.

1:6:0.6 concrete mix gauged at 20°C

3 metre pour

1.5 metre pour0.5 metre pour

60

50

40

30

20

10

Tem

pera

ture

(°C

)

0 20 40 60 80 100 120 140 160 180

Time (hours)

Cements 1/21

This is because the gypsum needles which have developed a structure in the paste arebroken and dispersed by re-mixing.

The cement manufacturer thus needs to optimize the level of dehydrated gypsum in thecement and match this to the reactivity of the C3A present. This concept is illustrated inFigure 1.12.

Figure 1.12 Optimization of soluble calcium sulfate.

Dry cement powder

Too much SO42–

in solution

Too little SO42–

in solution

On addition ofwater:

SO42– correct

Flash setC4AH13 or monosulfate(3C3A·CaSO4·12H2O)forms

Balanced retardationFine-grained ettringiteC3A·3CaSO4·32H2Oforms on C3A

False setCaSO4·2H2Oprecipitatesfrom solution

Many natural gypsums contain a proportion of the mineral natural anhydrite (CaSO4

– not to be confused with soluble anhydrite which is produced by gypsum dehydration).This form of calcium sulfate is unaffected by milling temperature and dissolves slowly inthe pore solution providing SO 4

2– ions necessary for strength optimization but having nopotential to produce false set.

Optimization of the level of readily soluble sulfate is achieved by a combination of:

• control of cement total SO3 level• controlling the level of natural anhydrite in the calcium sulfate used (either by purchasing

agreement or by blending materials)• controlling cement milling temperature

It must be emphasized that the optimum level of dehydrated gypsum can be influencedstrongly by the use of water-reducing admixtures. Cement–admixture interactions arecomplex and some admixtures will perform well with certain cements but may performrelatively poorly with others.

1.5.6 Techniques used to study hydration

The following techniques are used to study the hydration of cement:

C2S

C3S

C3A

C3A

C3A

C3A

C3A

C3A

C3A

C3A

C2S

C3SC3S

C2S

C3S

C2S

C3 S

1/22 Cements

• thermal analysis• X-ray diffraction• scanning electron microscopy

In order to study hydrated cement it is necessary to remove free water and arrest hydration.This is normally done by immersing the sample in acetone and drying it at low temperature(<50°C) in a partial vacuum.

Thermal analysisThe most commonly applied technique is thermo-gravimetric analysis. A small sample ofhydrated cement is placed in a thermobalance and the change in weight recorded as thesample is heated at a controlled rate from ambient to ~1000°C. The technique enables theproportion of certain hydrates present, such as ettringite and Ca(OH)2 to be determinedquantitatively.

X-ray diffractionThis technique is rapid but provides limited information as many of the hydrates present,notably C–S–H gel and calcium aluminate monosulfate are poorly crystalline and give ill-defined diffraction patterns.

Scanning electron microscopy (SEM)This is a powerful technique, particularly when the microscope is equipped with a microprobeanalyser. It involves techniques akin to X-ray fluorescence to determine the chemicalcomposition of hydrates in the field of view. The high resolution of the SEM enables themicrostructure of the hydrated cement paste in concrete or mortar to be studied. However,caution must be exercised when interpreting the images as specimen preparation and thevacuum required by most microscopes can generate features, which are not present in themoist paste.

1.5.7 Constitution of hydrated cement paste

Using the techniques described above it is possible to determine quantitatively the phasesformed as Portland cement hydrates (Taylor, Mohan and Moir, 1985). A typical exampleis given in Table 1.5. It can be seen that the dominant phase present (by volume) is C–S–H with approximately equal quantities of calcium hydroxide and monosulfate. Thesehydrate proportions are changed significantly when Portland cement is blended or intergroundwith pozzolanas (such as fly ash) or granulated blastfurnace slag. These reactions and thehydration products are discussed briefly in section 1.6.2.

1.6 Portland cement types

1.6.1 Standards

All developed countries have their own national standards for cements. These standardsdefine the permitted cement composition and set performance requirements for properties

Cements 1/23

such as setting time and development of compressive strength. They also describe the testprocedures to be used to determine cement composition and cement properties. Althoughit is now rather out of date, particularly in relation to the standards in place in Europeancountries, the publication by Cembureau, (1991) provides a useful review of cementtypes produced around the world.

In the past, national cement standards in many countries have been strongly influencedby those developed in the UK and published by the British Standards Institution (BSI)and by those developed in the USA and published by the American Society for Testingand Materials (ASTM). In 1991 The BSI published revised cement standards, which wereclosely aligned with the draft European Standard for Common Cements. This EuropeanStandard (EN 197-1) was adopted in 2000 by the following European countries: Austria,Belgium, Denmark, Finland, France, Germany, Greece, Iceland, Ireland, Italy, Luxembourg,Norway, Portugal, Spain, Sweden, Switzerland and the United Kingdom. The objective ofthis standard (in common with standards for other materials) is to remove barriers totrade. In order to meet this objective, existing national standards in the above countrieswere withdrawn in 2002. It can be expected that this European standard and the supportingstandards for test methods (EN 196) and for assessment of conformity (EN 197-2) willhave a strong influence on national (or regional) cement standards in the future.

1.6.2 Main cement types

Generic typesTable 1.6 summarizes the main generic types of Portland cement produced around theworld. The descriptions given in the table are very general. National (and regional)standards have specific limits for the contents of constituents and there may be several

Table 1.5 Constitution of hydrated Portland cement paste

Anhydrous cementWeight %

C3S 63C2S 13C3A 9.5C4AF 7.5CaSO4 5

Hydrated cement pasteVolume %

3 days 28 days 365 days

Unreacted C3S 18 8 5C2S 7 5 2C3A 4 0 0C4AF 5 5 2

Hydrates Calcium hydroxide [Ca(OH)2] 13 18 18C–S–H gel 39 47 54Monosulfate (C3A·CaSo4·12H2O 4 16 20Ettringite (C3A·3CaSO4·32H2O) 11 0 0

1/24 Cements

different ‘sub-types’ of cement with different levels of power station fly ash and blastfurnaceslag and even mixtures of slag, fly ash and limestone. In this chapter the term compositeis applied to all cements containing clinker replacement materials (other than a minoradditional constituent).

Types of ‘pure’ Portland cementTable 1.7 compares the typical chemical composition of normal grey, sulfate-resistingand white Portland cements. Many countries have national specifications for sulfate-resisting Portland cement (e.g. BS 4027, ASTM C150 Type V, French type CPA-ES). TheEuropean Committee responsible for cement standardization has been unable to reachagreement on the maximum C3A level required to ensure sulfate resistance and sulfate-resistant Portland cements are not recognized as a separate cement type in EN 197-1. Themaximum permitted C3A level (calculated by the Bogue method) in BS 4027 is 3.5% andin ASTM C 150 (Type V) 5%.

Sulfate-resisting clinker is normally produced in relatively short production runs by a

Table 1.6 Main types of cement produced around the world

Type Designation Constituents (excluding Applicationsminor additionalconstituent permitted insome countries)

Normal Portland ‘Pure’ Clinker and calcium sulfate All types of constructionexcept where exposed tosulfates

Sulfate-resisting ‘Pure’ Low C3A clinker and Where concrete is exposedPortland calcium sulfate to soluble sulfates

White Portland ‘Pure’ Special low iron content For architectural finishesclinker and calcium sulfate

Portland fly ash Composite Clinker, fly ash and calcium All types of construction.cement sulfate Some countries recognize

improved sulfate resistanceand protection against alkalisilica reaction if fly ashlevel above ~25%. Low heatproperties at high fly ash levels

Portland slag Composite Clinker, granulated All types of construction.cement blastfurnace slag and Most countries recognize

calcium sulfate improved sulfate resistanceif slag level above ~ 60%.Protection against alkali silicareaction recognized in somecountries. Low heat propertiesat high slag levels

Portland Composite Clinker, limestone of In Europe all types oflimestone specified purity and constructioncement calcium sulfate

Pozzolanic Composite Clinker, natural pozzolana All types of construction.cement and calcium sulfate Sulfate resistance properties

not normally recognized

Masonry cement Composite Clinker, limestone (or lime) Brickwork, blockwork andand air-entraining agent rendering

Cements 1/25

rotary kiln, which also produces conventional grey clinker. The frequent changeoversdisrupt the normal production process and have an adverse effect on the lifetime of therefractory bricks, which line the kiln.

Sulfate-resisting concrete can also be produced by ensuring an appropriate level of flyash or blastfurnace slag. This can be achieved either by purchasing a factory-producedcement or (where national provisions permit) by blending fly ash or ground slag withcement. The current UK codes (BS 8500: 2002 Part 2 ) require a minimum of 25% fly ashor 70% blastfurnace slag, by weight of the combination.

The production of white cement clinker requires careful selection of materials andfuels to ensure the minimum content of iron oxide and of other colouring oxides such aschromium, manganese and copper. In order to achieve the best possible colour the clinkeris normally fired under conditions where there is a slight deficiency of oxygen resultingin reduction of the colouring oxides to lower oxidation states, which have a lesser detrimentaleffect, and the clinker is quenched rapidly with water to prevent oxidation. All thesemeasures increase the production cost and white cement sells at a significant premiumover grey Portland cement.

Composite cementsComposite cements are cements in which a proportion of the Portland cement clinker isreplaced by industrial by-products, such as granulated blastfurnace slag (gbs) and powerstation fly ash (also known as pulverized-fuel ash or pfa), certain types of volcanicmaterial (natural pozzolanas) or limestone. The gbs, fly ash and natural pozzolanas reactwith the hydration products of the Portland cement, producing additional hydrates, whichmake a positive contribution to concrete strength development and durability. (Massazza,1988; Moranville-Regourd, 1988).

Table 1.7 Typical composition of grey, white and sulfate-resisting Portland cement

Grey Sulfate-resisting White% % %

SiO2 20.4 20.3 24.6Insoluble residue (IR) 0.6 0.4 0.07Al2O3 5.1 3.6 1.9Fe2O3 2.9 5.1 0.3CaO 64.8 64.3 69.1MgO 1.3 2.1 0.55SO3 2.7 2.2 2.1Na2O 0.11 0.10 0.14K2O 0.77 0.50 0.02Loss ignition (LOI) 1.3 1.3 0.85

Free lime % 1.5 2.0 2.3LSF % 96.6 97.3 95SR 2.55 2.33 11.2AR 1.76 0.71 6.3

Calculated mineralcompositionC3S 57 62 67C2S 16 12 20C3A 9 0.9 4.5C4AF 9 16 0.9

1/26 Cements

In contrast, finely ground limestone, while not hydraulically active, modifies the hydrationof the clinker minerals. It is introduced to assist in the control of cement strength developmentand workability characteristics. In some European countries, notably Belgium, France,the Netherlands and Spain, the quantity of composite cements produced considerablyexceeds that of ‘pure’ Portland cement. According to statistics supplied by the associationof European cement manufacturers (Cembureau) the average proportion of ‘pure’ Portlandcement delivered in European Union countries was 38% in 1999 the remainder being‘composite cements’. The proportion of composite cements in the UK is at present verymuch lower, and the UK also differs from most European countries in that the addition ofground granulated blastfurnace slag (ggbs) and fly ash at the concrete mixer is wellestablished. Further details are given in section 1.6.4.

Currently (2002), although there are regional variations, most of the ready-mixedconcrete produced in the UK contains either ggbs (at ~50% level) or fly ash (at~30% level).

The partial replacement of energy-intensive clinker by an industrial by-product or anaturally occurring material not only has environmental advantages but also has thepotential to produce concrete with improved properties including long-term durability.

The characteristics of the constituents of composite cements are summarized inTable 1.8.

Table 1.8 Nature of composite cement constituents

Constituents Siliceous fly ash Natural pozzolana Granulated Limestone(bituminous) blastfurnace slag

Reaction type Pozzolanic Latently hydraulic Hydration modifier

Brief Partly fused ash Material of volcanic Produced by rapid Calcium carbonatedescription from the combustion origin such as ash quenching of molten of specified purity

of pulverized coal blastfurnace slagin power stations

Typicalcomposition Range Range Range TypicalSiO2 38–64 60–75 30–37 3Al2O3 20–36 10–20 9–17 0.5Fe2O3 4–18 1–10 0.2–2 0.5CaO 1–10 1–5 34–45 51MgO 0.5–2 0.2–2 4–13 2S – – 0.5–2 –SO3 0.3–2.5 <1 0.05–0.2 0.3LOI 2–7 2–12 0.02–1 42K2O 0.4–4 1–6 0.3–1 0.1Na2O 0.2–1.5 0.5–4 0.2–1 0.02Reactive Low lime silicate Low lime silicate High lime silicate Calcium carbonatephases glass glass. Occasionally glass

zeolite type material

Pozzolanic materials contain reactive (usually in glassy form) silica and alumina,which are able to react with the calcium hydroxide released by hydrating cement, to yieldadditional C–S–H hydrate and calcium aluminate hydrates. This reaction is much slowerthan the hydration of the clinker silicates and the strength development of pozzolaniccements is slower than that of ‘pure’ Portland cements. Provided moist curing is maintained,

Cements 1/27

ultimate strengths may be higher than those of ‘pure’ Portland cement concrete. The C–S–H formed has a lower Ca:Si ratio than that found in pure Portland cement (~1.2 cf 1.7).

The nature of additions and the factors determining their performance are reviewed indetail in Chapter 3.

Fly ash has a significant advantage over natural pozzolanas as a result of the sphericalshape of the glassy particles. These normally have a positive influence on concrete workabilityenabling concrete water contents to be reduced and thus offsetting the early age strengthreduction. Fly ash performance may be improved by either removing coarse particles(using a classifier similar to that employed in closed-circuit cement grinding) or by co-grinding the ash with clinker. A British Standard (BS 3892: Part1) requires the ash to havea maximum 45 micron residue of 12% and ash meeting this standard is invariably classified.Blastfurnace slag is latently hydraulic and only requires activation by an alkaline environmentto generate C–S–H and calcium aluminate hydrates. Although strength development isslower than with ‘pure’ Portland cement it is practicable to produce cements (or concretes)containing 50% slag (by weight of binder) with the same 28-day strength. Early strengthsare considerably lower and those at 3 days may be approximately half those of ‘pure’Portland cement (at 20°C).

The ‘secondary’ hydrates from the pozzolanic and latent hydraulic reaction develop acement paste phase which has a lower permeability to water and to potentially aggressiveions such as sulfates and chlorides and this can have a positive impact on concretedurability.

While limestone constituents do not contribute significantly to strengths at 28 days,they do accelerate Portland cement hydration, and the reduction in early strength isnormally less than in the case of fly ash. The most important characteristic of a limestoneconstituent is that it should comply with the purity requirements of the relevant standard(BS EN 197-1). Portland limestone cements with levels of limestone as high as 30% havebeen produced in France and Italy for many years and are used in a wide range ofconstruction applications.

Figure 1.13 compares the strength development properties of cement prepared byblending a Portland cement with 20% ground limestone, granulated slag, natural pozzolanaand fly ash (Moir and Kelham, 1997).

In Figure 1.13 the levels of slag and fly ash are lower than those found in UK cementsor concretes but maintaining a constant addition level enables the reactivity of the materialsto be compared directly.

1.6.3 The European Standard for Common Cements (EN 197-1)

As described above, CEN member countries voted to adopt EN 197-1 in 2000. In 2002‘conflicting’ British Standards (such as BS 12) will be withdrawn. The British Standardfor sulfate-resisting cement, BS 4027, will continue until such time as agreement isreached on a European Standard for sulfate-resisting cement. Table 1.9 summarizes therange of cement compositions permitted by EN 197-1.

While these are ‘common cements’ they are not all available in all CEN membercountries. For example, Portland burnt shale cement requires a particular shale type,which is only found in southern Germany.

As well as defining the range of permitted cement compositions EN 197-1 also defines:

1/28 Cements

PC

Limestone

Slag

Pozzolana

Fly ash

1 10 100Age (days)

60.0

50.0

40.0

30.0

20.0

10.0

0.0

Com

pres

sive

str

engt

h (M

Pa)

Figure 1.13 Comparative strength growth at 20°C of cements containing 20% limestone, slag, fly ash andpozzolana (BS 4550 concrete, w/c 0.60).

Table 1.9 Cement types and compositions permitted by EN 197-1

Cement Notation Clinker Additiontype % %

CEM I Portland cement CEM I 95–100 –

CEM II Portland-slag cement II/A-S 80–94 6–20II/B-S 65–79 21–35

Portland-silica fume cement II/A-D 90–94 6–10Portland-pozzolana cement II/A-P 80–94 6–20

II/B-P 65–79 21–35II/A-Q 80–94 6–20II/B-Q 65–79 21–35

Portland-fly ash cement II/A-V 80–94 6–20II/B-V 65–79 21–35II/A-W 80–94 6–20II/B-W 65–79 21–35

Portland-burnt shale cement II/A-T 80–94 6–20II/B-T 65–79 21–35

Portland-limestone cement II/A-L 80–94 6–20II/B-L 65–79 21–35II/A-LL 80–94 6–20II/B-LL 65–79 21–35

Portland-composite cement II/A-M 80–94 6–20II/B-M 65–79 21–35

CEM III Blastfurnace cement III/A 35–64 36–65III/B 20–34 66–80III/C 5–19 81–95

CEM IV Pozzolanic cement IV/A 65–89 11–35IV/B 45–64 36–55

CEM V Composite cement V/A 40–64 36–60V/B 20–38 61–80

Notes:All cements may contain up to 5% minor additional constituent (mac)CEM V/A Composite cement contains 18–30% blastfurnace slagCEM V/B Composite cement contains 31–50% blastfurnace slagProportions are expressed as % of the cement nucleus (excludes calcium sulfate)

Cements 1/29

• the permitted chemical composition of the individual constituents• the permitted nature of the minor additional constituents (essentially either one of the

permitted main constituents or a material derived from the clinker-making process)• the permitted level of additives (e.g. grinding aids) – maximum of 1% in total including

less than 0.5% organic• minimum and maximum strengths for different strength classes• minimum setting time• chemical requirements (e.g. maximum SO3, MgO, loss on ignition (LOI), insoluble

residue (IR))• conformity criteria to demonstrate compliance with EN 197-1.

Table 1.10 summarizes the compositional requirements for the constituents of EN 197-1cements.

Table 1.10 Compositional requirements for constituents of EN 197-1 cements

Constituent Requirements

Clinker Minimum of 2/3rds C3S plus C2S. Ratio of CaO/SiO2 >2. MgO < 5%

Blastfurnace slag Minimum of 2/3rds CaO + MgO + SiO2. Ratio (CaO + MgO)/SiO2 > 1. Minimum2/3rds glass

Siliceous fly ash LOI < 7% (provided performance requirements for durability and admixturecompatibility are met, otherwise <5%). Reactive CaO < 10% and free CaO < 1%.Free lime up to 2.5% permitted if soundness test passed

Calcareous fly ash LOI < 7% (provided performance requirements for durability and admixturecompatibility are met, otherwise <5%). Reactive CaO >10%. If reactive CaO>10% and <15% then reactive SiO2 should be > 25%

Natural pozzolana Reactive SiO2 > 25%

Burnt shale Compressive strength of >25 MPa at 28 days when tested according to EN 196-1.Expansion of <10 mm when blended with 70% cement

Limestone CaCO3 > 75%. Methylene blue adsorption (clay content) < 1.2g/100g. Totalorganic carbon < 0.2% (LL) or 0.5% (L)

Silica fume LOI < 4%. Specific surface (BET) > 15.0 m2/g

The compositional requirements for blastfurnace slag are the same as those in theBritish Standard for Portland blastfurnace cement (BS 146) and for ground granulatedblastfurnace slag (BS 6699). Note that in EN 197-1 no test method is specified to determinethe minimum glass content of 2/3rds.

Similarly the requirements for fly ash are essentially the same as those in the BritishStandards for Portland pulverized-fuel ash cement (BS 6588) and for pulverized-fuel ash(BS 3892 Part 1) although there are minor differences related to maximum LOI andCaO content.

Standards for concrete additions are reviewed in greater detail in Chapter 3.Table 1.11 summarizes the mechanical and physical requirements for the strength

classes permitted by EN 197-1. Compressive strength is determined using the EN 196-1mortar prism procedure, which is outlined in section 1.7. Setting times are determined bythe almost universally applied Vicat needle procedure and soundness by the method firstdeveloped by Le Chatelier in the nineteenth century. These methods are described inEN 196-3.

1/30 Cements

Currently, in the UK (2002) all bulk cement supplied is class 42,5 or 52,5 and the vastmajority is pure Portland type (CEM I using BS EN 197-1 terminology). Less than 5% ofthe cement supplied is ‘composite’ and includes fly ash or limestone introduced duringmanufacture.

In contrast, in EU member countries (1999 data) the proportion of the three mainstrength classes was:

Class Production(% of total tonnage)

32,5 48%42,5 42%52,5 10%

The proportion of the different cement types was as follows:

Type Production(% of total tonnage)

CEM I Portland 38%CEM II Portland composite 49%CEM III Blastfurnace/slag 7%CEM IV Pozzolanic 5%CEM V Composite (and others) 1%

Thus, approximately 50% of the cement supplied was CEM II composite of which thehighest proportion was Portland limestone cement (40% of the CEM II and 20% of thetotal tonnage).

In the UK the established practice is to add ground-granulated slag (to BS 6699) orpulverized-fuel ash (to BS 3892 Part 1) direct to the concrete mixer and to claim equivalenceto factory-produced cement. The procedures required to demonstrate equivalence aredescribed in BS 5328, which will be replaced, by BS EN 206-1 and the UK complementarystandard, BS 8500 on 1 December 2003. In addition, some UK cement standards includecements with strength classes and properties outside the scope of BS EN 197-1 forcommon cements. For example, BS 146:2002, includes a low early strength class (L) forblastfurnace slag cements and there is a low 28-day class (22,5) in BS 6610:1996 for

Table 1.11 Strength setting time and soundness requirements in EN 197-1

Compressive strength MPa (EN 196-1 method)Strength Initial Soundnessclass Early strength Standard strength setting time (expansion)

(mm) (mm)2 days 7 days 28 days

32,5 N – ≥16,0 ≥32,5 ≤52,5 ≥7532,5 R ≥10,0 –

42,5 N ≥10,0 – ≥42,5 ≤62,5 ≥60 ≤1042,5 R ≥20,0 –

52,5 N ≥20,0 – ≥52,5 ≥4552,5 R ≥30,0 –

Cements 1/31

pozzolanic fly ash cement. Both of these standards will be withdrawn when EuropeanStandards covering the same scope are eventually published.

The chemical requirements of EN 197-1 cements are summarized in Table 1.12.

Table 1.12 Chemical requirements of EN 197-1 cements

Property Cement type Strength class Requirements

Loss on ignition CEM I (pure Portland) All ≤5.0%CEM III (>35% slag)

Insoluble residue CEM I

CEM III (.>35% slag All ≤5.0%

Sulfate (as SO3) 32,5 NCEM I 32,5 R ≤3.5%CEM II Portland composite cement 42,5 NCEM IV Pozzolanic cement 42,5 RCEM V Composite cement 52,5 N ≤4.0%

52,5 RCEM III (> 35% slag) All

Chloride All All ≤0.10%

Pozzolanicity CEM IV Pozzolanic cement All Satisfies the test(EN 196–5)

MgO All All ≤5.0% in clinker

Upper limits for loss on ignition (LOI) and insoluble residue (IR) have featured incement standards since their first introduction. The LOI ensured cement freshness and theIR limit prevented contamination by material other than calcium sulfate and clinker.However, the option to introduce up to 5% minor additional constituent (mac) has erodedtheir relevance. A higher level of assurance of consistency of performance is provided bythe much more rigorous performance tests, which must be performed on random despatchsamples at least twice per week.

The upper limit for SO3 features in all cement standards and its purpose is to preventexpansion caused by the formation of ettringite from unreacted C3A once the concretehas hardened. This expansive reaction, which occurs at normal curing temperatures a fewdays after mortar or concrete is mixed with water, should be distinguished from thephenomenon of delayed ettringite formation (DEF). All Portland cement can exhibitexpansion as a result of DEF if they are subjected to high initial curing temperatures(above 80°C) and subsequent moist storage conditions. The cement factors which increasethe risk of DEF have been identified by Kelham (1996). The cement SO3 level has apositive influence on cement strength development, particularly at early ages, and overthe past 20 years there has been a trend to raise the upper limit.

The purpose of the upper chloride limit is to reduce the risk of corrosion to embeddedsteel reinforcement. Although the upper limit is 0.1%, for prestressed concrete applicationsa lower limit may be agreed with the supplier and this will be declared on documentation.The concrete producer must, of course, consider all sources of chloride (water, aggregates,cement and admixtures) when meeting the upper limit for chloride in concrete.

The MgO limit of 5% on clinker ensures that unsoundness (expansion) will not occuras a result of the delayed hydration of free MgO. When MgO is present above ~2% inclinker it occurs as crystals of magnesium oxide (periclase) and these react relatively

1/32 Cements

slowly to form Mg(OH)2 (brucite) the formation of which is accompanied by expansion.Some countries such as the USA have a higher MgO limit (6%) but include an acceleratedexpansion test in which a sample of cement paste is heated in an autoclave and theexpansion must remain below 0.08%.

EN 197-1 also describes the testing frequencies and the method of data analysisrequired to demonstrate compliance with the requirements of the standard. Note that thevalues given in Tables 1.11 and 1.12 are not absolute limits. A given percentage of theresults obtained on random despatch samples may lie above or below these values. Forexample, for compressive strength 10% of results may lie above the upper limit forstrength in a particular strength class but only 5% below the lower limit. For the physicaland chemical requirement 10% of results may lie outside the limits. The spot samplestaken at the point of cement despatch are known as autocontrol samples and the testresults obtained as autocontrol test results.

Certificates confirming compliance with the requirements of the standards can beissued by EU Notified Certification Bodies (e.g. BSI Product Services) who follow theprocedures detailed in EN 197-2. As EN 197-1 is a harmonized standard, the certificationbody can issue EC certificates of conformity which permit the manufacturer to affix theCE marking to despatch documents and packaging. The CE marking indicates a presumptionof conformity to relevant EU health and safety legislation and permits the cement to beplaced on the single European market. Failure to consistently meet the requirements ofthe standard may result in withdrawal of the EC certificate.

1.6.4 Other European cement standards

The CEN committee responsible for the development of standards for cements and buildinglimes (CEN TC 51) has produced a number of standards for construction binders, whichare close to finalization and adoption by member countries these are:

prEN 413-1: Masonry cementprEN 645: Calcium aluminate cement (see Chapter 2)EN 197-1: prA1: Amendment to EN 197-1 to include low heat common cementsprEN 197-X: Low early strength blast furnace cementsprEN 14216: Very low-heat special cementsEN 459-1: Building lime (published)EN 13282: Hydraulic road binders (published)

Progress in developing a standard for sulfate-resisting cement is difficult as a result ofnational differences of view concerning the maximum level of C3A in CEM I cementsand the effectiveness of fly ash in preventing concrete deterioration. The solution may bea laboratory performance test but difficulties have been experienced in achieving a satisfactorylevel of reproducibility.

1.6.5 Other national standards

While the new European Standards can be expected to have a strong influence on thefuture development of national standards around the world, ASTM-based standards arelikely to remain important in many countries. The main differences between currentASTM cement standards and EN 197-1 are shown in Table 1.13.

Cements 1/33

1.7 Cement production quality control

Different customers have different priorities concerning the specific properties of greatestimportance to their business and Table 1.14 attempts to summarize these. The ‘cementproperty’ of prime importance is consistency of performance in relation to the aspects ofgreatest importance to the customer.

Table 1.13 Comparison between ASTM standards and EN 197-1

ASTM C 150 EN 197-1

Compressive strength Minimum values only Three main strength classes withupper and lower limits

Initial set Min 45 mins all types Class 32,5 min 75 minsClass 42,5 min 60 minsClass 52,5 min 45 mins

Minor additional Not permitted Permitted (<5%)constituent

SO3 Type I C3A ≤ 8% max 3.0% 32,5 and 42,5 N max 3.5%Type I C3A > 8% max 3.5% 42,5 R and 52,5 max 4.0%Slag and pozzolanic cements max 4.0%

Chloride No limit Max 0.1%

MgO Max 6.0% in cement Max 5.0% in clinker

Table 1.14 Cement quality aspects of importance to customers

Primary aspect Secondary aspects(of importance to most customers) (can be important to some customers)

Strength at 28 days Performance with slag and ashEarly strength Response to changes in w/c ratioWater demand Performance with admixturesStrength growth (and correlation between Premature stiffeningearly and late strengths) Delayed stiffeningAlkali content Heat of hydrationconsistency of these aspects Colour

FlowabilityTemperatureStorage propertiesConsistency of these aspects

While the cement manufacturer can control certain properties, such as alkali contentduring the manufacturing process it is not possible to control strength directly because ofthe time delay before results are obtained. In practice control over cement properties isachieved by control of

• raw mix chemistry, fineness and homogeneity• clinker chemistry and degree of heat treatment (burning)• cement fineness, SO3 level and forms of SO3 present.

The chemical composition of raw materials for cement making, clinkers and cementscan be determined by so-called wet chemical analysis methods. (BS EN 196-2, 1995).

1/34 Cements

This involves dissolving the materials in acid, and using a range of analytical techniqueswhich are specific to the elements to be determined. The analysis requires between 2 and6 hours to complete, depending on the facilities available, to determine the levels of theoxides S, A, F and C. Wet techniques also require highly skilled staff, if reliable resultsare to be obtained.

Fortunately, a rapid method of analysis known as X-ray fluorescence became availablein the late 1960s and this technique is almost universally applied at cement works aroundthe world. Sample preparation and the principles of the technique are outlined in Figure1.14.

Figure 1.14 Outline of X-ray fluorescence chemical analysis.

Sample of rawmeal or cement

Mix with fluxGrinding

millCrusher

Sampleof clinker

High-pressure

press

Fusionfurnace

Presseddisc

GlassbeadCrystal

X-ray tube

PrimaryX-rays Data

Detector

FluorescenceX-rays

Presseddisc orglass bead

The sample of raw meal, clinker or cement is inserted in the analyser, either in theform of a pressed disc of finely ground material or after fusing into a glass bead. In theanalyser, the specimen is irradiated with X-rays, which cause secondary radiation to beemitted from the sample. Each chemical element present emits radiation of a specificfrequency, and the intensity of the radiation is proportional to the quantity of that elementpresent in the sample.

All UK cement plants (and all modern cement plants around the world) are equippedwith an X-ray fluorescence analyser and use the technique to determine and control thelevels of CaO, SiO2, Al2O3, Fe2O3, MgO, K2O, Na2O, SO3 and Cl in cement raw meal andclinker. The current cost of manually operated equipment is ~£150 000.

Cements 1/35

There is a trend towards continuous on-line analysis of raw materials either before orjust after the raw mill using the technique of prompt gamma ray neutron activationanalysis (PGNAA), (Macfadyen, 2000). There is also a trend towards fully automatedlaboratories where the processes outlined in Figure 1.13 are undertaken by a robot,similar to those used, for example, in car assembly.

The level of uncombined lime (free lime) present in clinkers is normally determinedby extracting the CaO into hot ethylene glycol and titrating the solution with hydrochloricacid. Free lime can also be determined using the technique of X-ray diffraction andthis is finding increasing favour as it is easier to automate than the glycol extractionmethod.

Close control over cement milling is essential to ensure a product with consistentproperties. Although cement particle size distribution can be determined directly usinglaser diffraction techniques there have been difficulties with stability of the measurementsover a period of time and the main methods most commonly used on cement plants forfineness determination are surface area (SA) determined by air permeability and sieving.In the UK the cement surface area is determined by placing a precise weight of cementin a ‘Blaine’ cell, which is then compacted to a constant volume. The time taken for afixed volume of air to pass through the cell is a function of the cement surface area. Theproportion of coarse particles present, as indicated by the 45-micron (or for finely groundcements 32-micron) sieve residue has a greater influence on cement 28-day strength thanthe surface area and it is important to monitor this parameter closely. The technique mostcommonly use is air-jet sieving where a stream of air passes through the sieve, agitatingthe material above the sieve and greatly speeding up the passage of sub-sized particlesthrough the sieve. Instrumental techniques to measure cement fineness are improving andthe trend is towards ‘on-line’ sampling and analysis linked to automatic mill control.

In European plants, cement strengths are determined using the EN 196-1 mortar testprocedure. This utilizes a mortar, which consists of 3 parts (by weight) sand to 1 partcement. The dry sand, which is certified as meeting the requirements of the standard, issupplied in pre-weighed plastic bags, which simplifies the batching and mixing procedure.The w/c ratio is fixed at 0.50 for all cement types and the mortar is placed in a mould,which yields three prisms of dimensions 40 × 40 × 160 mm. Compaction is either byvibration or by ‘jolting’ on specially designed tables. The prisms can be broken in flexureprior to compressive strength testing but this stage is normally omitted and the prismssimply snapped in two using a simple, manually operated, breaking device. The brokenends of each prism are tested for compressive strength using platens of dimensions 40 ×40 mm. Normally three prisms are broken at each test age and the result reported is themean of six tests. A typical class 42,5 cement will give a strength of 55–59 MPa at 28days. This is a similar strength level to that obtained if the cement is tested in a crushedgranite concrete (such as the materials used in BS 4550 testing) with a w/c of 0.50. MostUK cement plants should achieve an annual average standard deviation for the 28-daystrength results of main products in the range 1.5–2.0 MPa.

The setting time of cement paste is determined using the Vicat needle according to EN196-3. This requires frequent testing of penetration in order not to ‘miss’ the initial andfinal set. Consequently there is a trend to introduce automation and equipment can bepurchased which will determine the setting time of a number of cement samples (e.g. 12)simultaneously.

Effective quality control must start with the quarrying, proportioning and blending of

1/36 Cements

the raw materials. If the raw mix is variable then no amount of adjustments to subsequentcontrol parameters will eliminate product variability.

1.8 Influence of cement quality control parameters onproperties

1.8.1 Key parameters

The cement properties of water demand (workability), setting behaviour and strengthdevelopment are largely determined by the following ‘key’ cement quality control parameters:

• Cement fineness (SA and 45 micron residue)• Loss on ignition (LOI)• Clinker alkalis and SO3

• Clinker-free lime• Clinker compound composition (mainly calculated C3S and C3A levels)• Cement SO3 and the forms of SO3 present

All of the above, with the exception of the forms of SO3 present, can be determinedrapidly by a competent cement plant laboratory. Specialist thermal analysis equipment isrequired to determine the forms of calcium sulfate present. Although some cement plantshave this equipment it is more generally found at central laboratories.

1.8.2 Water demand (workability)

The initial hydration reactions and the importance of matching the supply of readilysoluble calcium sulfate to clinker reactivity were reviewed in detail in section 1.5.5. Inorder to ensure satisfactory workability characteristics the cement producer should:

• maintain an appropriate ratio of SO3 to alkalis in the clinker• achieve a uniform and appropriate level of dehydrated gypsum by a combination of

controlling the level of natural anhydrite in the ‘calcium sulfate’ used and the cementmilling conditions

The partial replacement of clinker by fly ash should result in a reduction in water demandwith a consequent increase in strength which partially offsets the lower early strengthsassociated with the lower reactivity of the ash. The reduced water demand can be mainlyattributed to the glassy spheres present in fly ash which lubricate the mix. The lowerdensity of the ash compared to cement also increases the cement paste volume. Blastfurnaceslag grinds to yield angular particles. The slag is unreactive during initial hydration andgenerally has a neutral influence on water demand. Limestone can have a positive influenceon water demand particularly when compared to a more coarsely ground pure Portlandcement of the same strength class. This is because the fine limestone particles result in amore progressive (optimized) cement particle size distribution with a lower proportion ofvoids, which must be filled with water.

Cements 1/37

1.8.3 Setting time

Cement paste setting behaviour is strongly influenced by clinker reactivity and by cementfineness. The main factors, which tend to shorten setting time, are increases in the levelsof:

• clinker-free lime• cement fineness (surface area SA)• C3S content• C3A content

Because the setting time test is carried out on a sample of cement paste, which is gaugedwith water to give a standard consistency, any change in cement properties, which increasesthe water required, for example a steeper particle size distribution, will extend the settingtime. The effects of a steeper size distribution are much more apparent in paste than inconcrete and concretes made from the same cement may not exhibit the same magnitudeof extension in setting time.

Setting time may also be extended by the presence of certain minor constituents. Anexample is fluorine. An increase in clinker fluorine level of 0.1% can be expected toincrease setting time by ~ 60 minutes.

Both fly ash and slag will increase setting time while a Portland limestone cement mayhave a slightly shorter setting time than the corresponding pure Portland cement.

1.8.4 Strength development

FinenessAs discussed in section 1.7, on cement plants, cement fineness is normally determinedwith reference to surface area (SA) and 45-micron residue. While surface area is a goodguide to the early rate of hydration of cement and thus early strengths, it is a less reliableguide to late strengths and, in particular, to 28-day strengths. This is because understandard curing conditions clinker particles which are coarser than approximately 30microns are incompletely hydrated at 28 days. For a given SA the lower the 45-micronresidue, the higher the 28-day strength. In EN 196-1 mortar an increase in 45-micronresidue level of 1.0% can be expected to lower 28-day strength by ~ 0.4 MPa. In general,more modern milling installations, and in particular those with high efficiency separation,will yield cements with steeper size distributions. The steeper size distribution mayrequire the introduction of a minor additional constituent (mac or filler) to control 28-daystrength and thus remain within a certain strength class (Moir, 1994).

Loss on ignition (LOI)Cements which contain up to 5% limestone minor additional constituent will have relativelyhigh LOI values as a result of the decomposition of CaCO3 during the test. In the case ofcements which do not contain a limestone mac an increase in LOI above the ‘base level’originating from residual water combined with calcium sulfate represents water and CO2,which have combined with the clinker during storage or cement milling. Note that insome cases the base level may be slightly increased by CaCO3 present as an impurity inthe calcium sulfate.

1/38 Cements

The pre-hydration of clinker represented by the increased LOI has a marked influenceon strength development. Typically, for a cement without a calcareous mac an increase inLOI of 1% can be expected to reduce EN 196-1 mortar 28-day strength by ~4 MPa.

The influence of LOI when a cement contains a calcareous mac is much less clear. Thereduction in strength accompanying an increase in LOI of 1% associated with a higherfiller level (1% LOI as CO2 represents an addition of 2.5% high-grade CaCO3) will resultin a 28-day strength reduction of ~1 MPa.

In cement with several sources of LOI more sophisticated techniques (such as X-raydiffraction, or direct determination of carbon) can be used to determine the CaCO3 level.

Clinker alkalis and SO3As described in section 1.3.6, the alkali metal oxides (Na2O and K2O) have a strongaffinity for SO3 and if sufficient SO3 is present they will crystallize as alkali sulfates. Inthis form they are readily soluble and quickly dissolve in the gauging water modifyingstrength development properties. Soluble alkalis accelerate early strength developmentand depress late (28-day) strengths (Oesbaek, Jons, 1980).

Almost all CEM I type cements available in Europe will have eqNa2O (calculated asNa2O + 0.66K2O) in the range 0.3–1.0%. Figure 1.15 illustrates the influence of cementalkalis (expressed as eqNa2O) on the strength development of EN 196-1 mortar prisms.Note that in these clinkers there was sufficient SO3 to combine most of the alkali presentas sulfates.

Figure 1.15 Influence of cement alkalis on strength development when present as soluble alkali sulfates.

70

60

50

40

30

20

10

0

MP

a

0.2 0.4 0.6 0.8 1.0

eqNa2O %

Note: Ratio of clinker SO3 to eqNa2O > 1.3

EN 196-1 mortar prisms

28 days

7 days

2 days

1 day

A similar reduction in late strength can be achieved by adding alkali sulfates to cement;the increase in early strengths is normally less than when the sulfates are present inclinker. At early ages the dissolved alkali sulfate accelerates C3S hydration. Afterapproximately 1 day of hydration the supply of sulfate from calcium sulfate is exhaustedand the pore solution consists essentially of Na+, K+ and OH– ions whose pH is a functionof the dissolved alkali. With high alkali cements the pH level may exceed 13.5 and thisappears to be associated with inhibited silicate hydration.

In EN 196-1 mortar an increase in cement eqNa2O (in soluble form) can be expectedto increase 2-day strength by ~ 0.8 MPa and reduce 28-day strength by ~1.5 MPa.

Data for experimental clinker prepared with a very high level of clinker alkali and SO3

Cements 1/39

show that the depression of 28-day strength remains almost linear but the acceleration ofearly strength reaches a maximum and a negative influence may be detected at very highlevels. Figure 1.16 illustrates the results obtained. Note that eqNa2O levels above 1% areunlikely to be encountered in Europe.

Figure 1.16 Influence of very high level of eqNa2O combined in clinker as alkali sulfate.

2d7d28d91d365d

EN 196-1 mortar

0 0.5 1 1.5 2 2.5eqNa2O %

80

70

60

50

40

30

20

10

0

Mor

tar

stre

ngth

(M

Pa)

If the level of clinker SO3 is insufficient to combine with the alkali then the acceleratingeffect is reduced but the depression of 28-day strength is almost the same. This is becausethe alkali held in solid solution in the clinker minerals is released as hydration progresses.

Cements produced from high alkali clinker tend to give a better performance with slagand fly ash than equivalent low-alkali clinker. This is partly because the pH of the poresolution is reduced both by dilution and by absorption of alkalis in the lower calcium C–S–H formed (Taylor, 1987) but also because of the more aggressive attack on the glassyphases present in the slag and fly ash.

Free limeCement-free lime levels should normally lie in the range 0.5–3%. Late strengths arenormally maximized by a low free lime level (as this maximizes the combined silicatecontent) but, as discussed in section 1.8.2, low free lime levels are associated with lowreactivity and extended setting times. For cements with ‘normal’ parameters of LSF ~95and SR 2.5 an increase of free lime level of 1% can be expected to reduce EN 196-1mortar 28-day strength by ~1.5 MPa.

Compound compositionStrengths at ages of 1, 2 and 7 days are almost linearly related to C3S content. At 28 daysC2S makes a significant contribution but the contribution does depend on C2S reactivity,which is determined by clinker microstructure and by impurities present in the crystallattice (solid solution effects). For example, the presence of belite clusters associated withcoarse silica in the raw mix will reduce the contribution from C2S at 28 days.

When clinkers were prepared in a batch rotary kiln with identical chemistries otherthan the ratio of C3S to C2S (Kelham and Moir, 1992) it was shown that with well-

1/40 Cements

combined clinkers the highest 28-day strength was obtained with a moderate C3S level.Results are illustrated in Figure 1.17.

Figure 1.17 Influence of C3S content on concrete strength development.

MPa60

50

40

30

20

10

025 35 45 55 65

C3S content

BS 4550 concrete cured at 20°C

1 day 3 days 7 days

28 days 91 days 365 days

At 2 days and 7 days an increase in cement C3S content of 10% can be expected toincrease EN 196-1 mortar 2-day and 7-day strengths by ~3.5 and ~5 MPa respectively.The influence on 28-day strength is less certain but will normally be slightly positive.

For C3A contents in the range 5–12% an increase in C3A level will normally increase28-day strength in standard quality control tests. This effect is more likely to be seen ifthe C3A content increases as a result of an increase in AR rather than a decrease in SR.A decrease in SR will also reduce the total silicate content. The positive influence of C3Ais attributed to its higher reactivity compared to C4AF, which results in a greater volumeof hydrates and thus lower cement paste porosity at 28 days.

In field concretes (as opposed to fixed w/c laboratory mortars) the benefits of a higherC3A content may be offset by a greater rate of slump loss resulting in a higher w/c ratioin order to maintain the required slump level.

SO3 level and forms of SO3 presentWhile the forms of calcium sulfate present can have a marked influence on the waterdemand of cement in concrete it is the total cement SO3, which has the primary influenceon strength properties. Anhydrite, if present, will dissolve during the first 24 hours ofhydration and be available to participate in the strength-forming hydration reactions inthe same manner as calcium sulfate from gypsum.

The response of a cement to a change in cement SO3 level is influenced by a numberof factors, which include:

• the alkali content and in particular the alkali sulfate (soluble alkali) content• the C3A level• the cement fineness.

Cements 1/41

When optimizing the SO3 level it is important to monitor the influence of the change onconcrete water demand/rheology. It is also important to ensure that cements with differentSO3 levels are compared on a meaningful basis. For example, at constant SA the higherSO3 cements will normally have higher 45-micron residues and consequently lower 28-day strengths.

Most clinkers will show a significant increase in early strength when the cement SO3

level is increased from 2.5% to 3.5%. The influence on 28-day strength is generally muchless but still positive. Typical results from laboratory tests are shown in Figure 1.18. In theexample shown the adverse influence of cement SO3 level on concrete slump could havebeen reduced by replacing a proportion of the gypsum by natural anhydrite.

Figure 1.18 Example of influence of cement SO3 level on concrete strengths.

Slump(mm) Cements ground in laboratory

ball mill at 120°C forconstant grinding time

2 2.5 3 3.5

60

40

20

100 mm cubes according to BS 4550

28 days

7 days

3 days

1 day

Compressivestrength(MPa)

40

30

20

10

2 2.5 3 3.5

SO3(%)

In most countries the opportunity to optimize cement SO3 is restricted by the upperlimits for SO3 in the relevant standards.

One important difference between a factory-produced composite cement and a concretemixer blend is that in the latter it is not possible to control SO3 level. With high slagblends in particular, the cement SO3 level will be much lower than that of a factory-produced cement.

1/42 Cements

1.9 Relationship between laboratory mortar results andfield concrete

Although UK cement producers replaced the BS 4550 concrete test with the EN 196-1mortar test in 1991 as the primary assessment of cement strength (as required by BS 12:1991) parallel testing in concrete has continued. A large quantity of data have beengenerated concerning the relationship between mortar and concrete strengths (Moir, 1999).Figure 1.19 shows the relationship between strength results obtained by the two methods.

Figure 1.19 Relationship between EN 196-1 mortar and BS 4550 concrete strengths.

PC 42.5SR 42.5

PC 52.5

Compressive strength EN 196-1 mortar (MPa)

70

60

50

40

30

20

1010 20 30 40 50 60

Each point represents themean of 23 tests oncements sampled over a4-year period

Compressive strength BS 4550 concrete (MPa)

Although there is a clear ‘average’ relationship it is also clear that different cementshave slightly different mortar to concrete (m/c) ratios and that this is a function of cementfineness and chemistry. This is not surprising as different cements respond differently tochanges in cement content and w/c. This phenomenon is illustrated in Figure 1.20.

It can be seen that the 28-day strength ranking of the cements was different at the threedifferent cement contents (and consequent w/c ratios). Thus no single test procedure,whether mortar or concrete, can reliably predict cement performance across a range ofcement contents. The most important influence is cement alkali content as low alkalicement will tend to give relatively higher 28-day strengths at low w/c ratios and highalkali cements will tend to give lower strengths. Size distribution is also important as athigh w/c ratios cement hydration will be more complete and the influence of coarseparticles becomes more important.

1.10 Applications for different cement types

A detailed review of cement applications and the specification of concrete are outside thescope of this chapter. However, Table 1.15 summarizes the main applications likely to beselected for the principal generic cement types (and mixer blend equivalents) available inthe UK. Note that this list has been prepared in 2002 and anticipates the publication of BS

Cements 1/43

EN 413-1. Additional cement types (drawn from the options available in BS EN 197-1)may become commercially significant in the future.

The choice of cement for structural concrete should be in accordance with BS 8500:2002 after implementation on 1 December 2003 when it will supersede BS 5328. BS 8500recognizes the sulfate-resisting properties of slag and fly ash concretes and the reducedrisk of asr if appropriate levels of these materials are present.

It must be recognized that in many applications that cement choice has a much lesserinfluence on concrete long-term performance than the practical aspects of mix control,cement content, water content, aggregate quality, compaction, finishing and curing.

1.11 Health and safety aspects of cement use

Dry cement normally has no harmful effect on dry skin. However, when cement is mixedwith water a highly alkaline solution is produced. The pH quickly exceeds 13 and ishighly corrosive to skin.

Precautions should be taken to avoid dry cement entering the eyes, mouth and nose

BS 4550Plant EN 196-1Plant ASTMC 109Plant

ABCDEFGHIJKLMNOPQ

32 34 36 38 40 42 44 46MPa

ABCDEFGHIJKLMNOPQ

46 48 50 52 54 56 58 60 62MPa

ABCDEFGHIJKLMNOPQ

36 38 40 42 44 46 48 50MPa

ABCDEFGHIJKLMNOPQ

6 8 10 12 14 16 18MPa

200 kg/m3 (60 mm slump)Plant

ABCDEFGHIJKLMNOPQ

60 65 70 75 80 85 90MPa

ABCDEFGHIJKLMNOPQ

32 34 36 38 40 42 44 46MPa

310 kg/m3 (60 mm slump)Plant

550 kg/m3 (60 mm slump)Plant

Figure 1.20 Influence of test procedure and w/c ratio on 28-day strengths.

1/44 Cements

and skin contact with wet cement, mortar or concrete should be avoided. Serious skinburns requiring skin grafts have occurred where concrete has penetrated workers’ bootsor where workers have kneeled for a prolonged period in fresh concrete wearing fabrictrousers.

In addition to the acute burns described above exposure to moist mortar or concreteover a period of time may result in irritant contact dermatitis. Some people are sensitive(or after several years’ exposure will become sensitized) to the water-soluble chromatepresent in cement. In Scandinavian countries ferrous sulfate has been added to cementduring manufacture to precipitate the soluble hexavalent chromium as the insoluble trivalentform. The maximum permitted level of water-soluble chromium in these countries is

Table 1.15 Applications for Portland cement types available in the UK

Cement BS BS EN 196-1 Clinker Main applicationstype designation content*

Portland BS CEM I 42,5 & 95–100 All types of general construction wherecement EN 52,5 N resistance to sulfates or control of

197-1 concrete temperature rise is not required.

Portland BS CEM I 52,5 R 95–100 As for 42.5 but where higher earlycement EN strengths are required (for example, in

197-1 precast concrete production or winter(concreting)

Packed BS CEM I 32,5 95–100Portland EN Where improved concrete placing andcement with 197-1 finishing properties and resistance tointegral air- freeze–thaw cycling are required withoutentraining the need to use an air-entraining agentagent

Sulfate- BS Class 42.5 100 Where resistance to sulfates present inresisting 4027 ground water is required. Alkali contentcement <0.60% (optional) and can be used to

reduce risk of alkali silica reaction

Masonry BS MC 22,5 ≥40 For brick laying mortar and rendering.cement EN Gives improved plasticity and adhesion

413-1 as well as improved water retention anddurability

Portland fly BS CEM II/B-V 65–79 All types of general construction. Whenash cement EN fly ash level exceeds 25% gives sulfate

197-1 resistance and protection against alkalisilica reaction. Moderate low heat at upperend of fly ash range (max 35%)

Pozzolanic BS CEM IV/B2 45–64 For grouts and low-heat applicationscement EN

197-1

Blastfurnace BS CEM IIIA 35–64 All types of general construction. Wherecement EN slag level above 40% gives protection

197-1 against alkali silica reaction. Moderatelow-heat properties

Blastfurnace BS CEM IIIB 20–34 Where sulfate resistance and/or low heatcement EN properties required

197-1

Portland BS CEM II/A-L 80–94 For all types of general constructionlimestone EN CEM II/A-LL except where resistance to sulfates iscement 196-1 required

*As percentage of nucleus, i.e. excluding gypsum.

Cements 1/45

2 ppm. Although the clinical evidence for the effectiveness of this measure is inconclusivethere are moves in Europe to adopt this approach more widely, at least in bagged if notbulk cement. The addition of ferrous sulfate has an influence on concrete rheologicalproperties and also concrete colour. In the UK, the manufacturers have decided to monitorthe level of soluble chromium and minimize levels by careful selection of raw materials.

References

Bhatty, J.I. (1995) Role of Minor Elements in Cement Manufacture and Use. Portland CementAssociation, Skokie, PCA RD 109.2T.

Blezard, R. (1998) History of calcareous cements. In Hewlett, P.C. (ed.), Lea’s Chemistry ofCement and Concrete. Arnold, London, pp. 1–19.

Bogue, R.H. (1955) The Chemistry of Portland Cement (2nd edn). Reinhold, New York.BS 8500: (2002) Concrete – Complementary British Standard to BS EN 206-1. Part 2: Complementary

requirements for constituent materials, designed and prescribed concrete, production and conformity.BSI, London.

BS EN 196-2: (1995) Methods of testing cement. Part 2. Chemical analysis of cement.Cembureau (1991) Cement Standards of the World (8th edn). Cembureau, Brussels.Coole, M.C. (1988) Heat release characteristics of concrete containing ground granulated slag in

simulated large pours. Concrete Research 40(144), 152–158.Kelham, S. (1996) The effect of cement composition and fineness on expansion associated with

delayed ettringite formation. Cement and Concrete Composites 18, 171–179.Kelham, S. and Moir, G.K. (1992) In 9th International Congress on the Chemistry of Cement, Vol.

5. NCCBM, Delhi, pp. 3–8.Killoh, D.C. (1988) A comparison of conduction calorimetry and heat of solution methods for

measurement of the heat of hydration of cement. Advances in Cement Research 1(3), 180–186.Macfadyen, J.D. (2000) Online chemical analysis for control. World Cement 13, 8, 70–73.Massazza F. (1988) Pozzolana and pozzolanic cements. In Hewlett, P.C. (ed.), Lea’s Chemistry of

Cement and Concrete. Arnold, London, pp. 471–631.Moir, G.K. (1994) Minor additional constituents. In Dhir, K.D. and Jones, M.R. (eds). Euro-

Cements Impact of ENV 197 on Concrete Construction, E&FN Spon, London, pp. 37–56.Moir, G.K. (1999) Factors influencing the ratio of mortar to concrete strengths. In Dhir and Jones

(eds), Proceedings of Conference on Creating with Concrete, Volume, Modern Concrete Materials:Binders Additives & Admixtures. Thomas Dyer, Dundee.

Moir, G.K. and Glasser, F.P. (1992) Proceedings of the 9th International Congress on the Chemistryof Cement, Vol. 1. NBC, New Delhi, pp. 125–152.

Moir, G.K. and Kelham, S.K. (1997) Developments in the manufacture and use of Portland limestonecements. Proceedings of ACI International Conference, Malaysia. SP 172-42, pp. 797–819.

Moranville-Regourd, M. (1988) Cements made from blastfurnace slag. In Hewlett, P.C. (ed.), Lea’sChemistry of Cement and Concrete, Arnold, London, pp. 633–674.

Odler, I. (1998) Hydration, setting and hardening of Portland cement. In Hewlett, P.C. (ed.), Lea’sChemistry of Cement and Concrete. Arnold, London, pp. 260–263.

Oesbaek, B. and Jons, E.S. (1980) The influence of the content and distribution of alkalis on thehydration properties of Portland cement. Proceedings of 8th International Congress on theChemistry of Cement. Septima, Paris, pp. II 135–140.

Taylor, H.F.W. (1987) A method for predicting alkali concentration in cement pore solutions.Advances in Cement Research, 1(1), 5–17.

Taylor, H.F.W. (1997) Cement Chemistry, Thomas Telford, London, pp. 113–225.Taylor, H.F.W., Mohan, K. and Moir, G. (1985) Analytical study of pure and extended Portland

cement pastes: I, Pure Portland cement pastes. Journal of the American Ceramic Society 68(12),680–685.

2

Calcium aluminatecementsKaren Scrivener

2.1 Introduction

Calcium aluminate cements (CACs) are the most important type of non-Portland orspecial cements. Even so, the volume used each year is only about one thousandth of thatof Portland cement. As they are considerably more expensive (four to five times), it istherefore not economic to use them as a simple substitute for Portland cement. Insteadtheir use is justified in cases where they bring special properties to a concrete or mortar,either as the main binder phase or as one component of a mixed binder phase.

Some of the properties that can be achieved through the use of CACs are:

• rapid hardening• resistance to high temperatures and temperature changes• resistance to chemical attack, particularly acids• resistance to impact and abrasion.

The uses of calcium aluminate cements today are extremely diverse. The largest singleuse is in refractory concretes, which is described in a separate chapter. CACs are alsoextensively used in combination with other mineral binders (e.g. Portland cement, calciumsulfate, lime) and admixtures to produce a range of specialist mortars for applicationssuch as repair, floor levelling, tile adhesives and grouts.

Uses in traditional concrete are generally found where severe conditions must befaced, such as extreme temperatures, exposure to aggressive chemicals, mechanical abrasionor impact, often in industrial situations and where these severe conditions are combined

2/2 Calcium aluminate cements

with a need for rapid return to service. Some typical applications are illustrated inFigure 2.1.

This chapter aims to give the reader a broad understanding of the properties andbehaviour of calcium aluminate cements as relevant to their present-day applications inorder to support appropriate and intelligent use of these materials. The aim has been forclarity, focusing on the essentials with a minimum of references. For further reading and

Figure 2.1 Some typical applications of CAC concrete. From top left:• Bridge widening, rapid hardening• Dam flushing gate, abrasion resistance• Foundry floor, abrasion and thermal shock resistance• Cyrogenic facility, thermal shock, freeze–thaw• Sewer, resistance to bacteriogenic acid corrosion• Fire training building, resistance to thermal shock

Calcium aluminate cements 2/3

details the reader is referred to Scrivener and Capmas (1998), Concrete Society (1997)and Mangabhai and Glasser (2001).

2.1.1 Terminology

As this is often a matter of confusion a brief explanation is given of some of the cementchemistry terms used.

A phase is the name given to distinct materials with a specific combination of elementswith a specific arrangement, or crystal structure. In all the phases found in cements,elements such as calcium, silicon, aluminium and iron are always present with a fixedquantity of oxygen. This leads to the use of a shorthand notation, where letters are usedto represent the fixed combinations of elements and oxygen or oxides:

C = CaO (lime)S = SiO2 (silica)A = Al2O3 (alumina)F = Fe2O3 ( ferric oxide)f = FeO ( ferrous oxide)$ = SO3 (generally referred to as sulfate)H = H2O (water)

Cements are made by heating the raw materials at high temperatures to give a clinker,which is a combination of several different phases. As there is no water (or hydrogenoxide) present in these phases they are sometimes referred to as anhydrous phases. Afterreaction with water, new phases form, in which water is chemically combined with theother oxides. These phases are known as hydrates.

Some cement phases are also commonly referred to by their mineral name these aregiven on first mention of the phase for completeness and as information for furtherreading.

2.2 Chemistry and mineralogy of CACs

The original motivation for the development of CACs was to find new cement chemistriesthat would be more resistant to sulfate attack than Portland cement. This led variousresearch groups to develop cements based on calcium aluminates and eventually to thepatenting of a viable method of large-scale production by Bied of the Lafarge Companyin 1908. Following this, other special properties, such as rapid hardening, were quicklydiscovered and used for a range of applications. For instance, in the First World War theseproperties were used in battlefields to build tank defences. However, calcium aluminatecements were not produced commercially until the 1920s.

Calcium aluminate cements, like Portland cements, contain the oxides of calcium,silicon, aluminium and iron. However, their composition is quite distinct as can be seenin Figure 2.2. This shows the approximate zone of compositions of calcium aluminatecements, within the CaO–SiO2–Al2O3 system along with the composition zones of Portlandcement and of blastfurnace slag.

Due to the requirements for refractory concretes, CACs are produced with a wide

2/4 Calcium aluminate cements

range of alumina (Al2O3) contents from around 40–80%. The higher alumina contentgrades are made from relatively pure raw materials (metallurgical alumina and lime),which make them very expensive. They are made in rotary kilns, similar to, but smallerthan, typical Portland cement kilns. These grades are used almost exclusively for refractoryconcretes, which are dealt with in a separate chapter (although they have the same basicchemistry of hydration, etc.).

Here we are mainly concerned with the lower alumina grades (~40–60% Al2O3),which are generally made by complete melting of bauxite and limestone in a ‘reverbatory’furnace*. The type of bauxite used determines the minor elements present and hence thecolour of the cement. Cements made with red bauxite contain about 20% of iron oxidesand have a brown, dark grey or black colour (for example, Ciment Fondu®†). Cementsmade with white bauxite contain little or no iron and are light grey to white. However,they still contain around 3–8% of silica (for example, Secar 51®†).

For these grades of CAC with lower alumina content the mineralogy (i.e. phasespresent) of the unreacted cement can be fairly complex, but the common feature of these,and all calcium aluminate cements, is the presence of monocalcium aluminate – CA orCaAl2O4. This is the principal reactive phase responsible for properties of the materialand typically makes up from 40% to 60% of the cement. The other reactive phase is C12A7

(mayenite). This phase is important because it helps to trigger the setting process. However,its presence is strictly controlled by the manufacturers since too much would lead to earlystiffening.

The other phases – C2S (belite), C2AS (gehlenite), spinel phase, ferrite solid solution,etc. (Touzo et al., 2001) – contain silicon and /or iron and other minor elements inaddition to calcium and aluminium. The combined quantities of these phases may makeup over half of the cement. Therefore they are important from the point of view of

* In a reverbatory kiln, the raw materials in the form of large lumps slowly descend a vertical stackwhere heat exchange with the exhaust gases occurs. At the base of the stack is the hearth, wherecomplete melting of the raw materials occurs. The molten product is tapped off continuously andsolidifies in large ingots.†Products of Lafarge Aluminates.

SiO2

Slags

Portlandcements Calcium

aluminatecements

Al2O3CaO

Figure 2.2 Composition range of calcium aluminate cements compared to Portland cements.

Calcium aluminate cements 2/5

manufacture and correct control of the cement compositions, but these phases have muchlower reactivity and play a negligible role in the initial hydration reactions, although theymay partially react at later ages or higher temperatures. Figure 2.3 shows the microstructure

Figure 2.3 Microstructure of clinker of a standard grade CAC.

of a standard grade calcium aluminate cement, with the most important phases labelled.The size of the dark grey areas of monocalcium aluminate varies from grain to grain(Figure 2.4), due to the different cooling rate of the cement as it solidifies in the ingots,

Figure 2.4 Grains of CAC, showing different size of crystallization.

1

6

5 microns

7

CA

‘Ferrite’

Pléochroïte

Spinel

Gehlenite

C2S

2/6 Calcium aluminate cements

after emerging as a melt from the furnace. The fine-textured regions come from the outerparts of the ingot, where the cooling rate is faster, and the coarser-textured regions fromthe centre, which cools more slowly. Despite the difference in the size of the areas, thechemical composition of the individual phases is very constant across the ingot. The areasof monocalcium aluminate are not individual grains, but are interconnected in threedimensions, so all of this phase is able to react without restriction by unreactive phases.

For standard grade CACs both the chemical content of alumina (Al2O3) and the contentof the phase, monocalcium aluminate (CA) lie in the range 40–60%, but there is no directequivalence between these two quantities. Most of the different phases contain aluminain differing amounts, and conversely CA, may contain small amounts of other elements(e.g. iron). The overall chemistry of a cement depends on the proportions and compositionof the raw materials and can be measured either by traditional wet chemistry techniquesor, more normally nowadays, by X-ray fluorescence (XRF). The phase composition ormineralogy is determined by this chemistry of the raw materials and thermodynamics. Itmay also be influenced by the parameters of cement production, e.g. furnace oxidizingconditions, cooling rate. However, the complexity of the mineralogy and the compositionsof the individual phases mean that it is not possible to derive simplified relationships,equivalent to the Bogue calculation for Portland cements. Quantitative analysis of thephases present is now possible by X-ray diffraction combined with Rietveld analysis, butcareful calibration and a good understanding of crystallography are necessary to obtainreliable results (Füllmann et al., 1999).

2.2.1 Basics of hydration and conversion

The chemical process by which cements react with water to give a rigid solid is knownas hydration. In terms of its physical process this is broadly the same for calcium aluminatecements as for Portland cements – namely anhydrous cement and water are replaced byhydrate phases, which have a larger solid volume and bridge the spaces between thecement grains. However, the products of calcium aluminate cement are chemically quitedifferent from those of the calcium silicate phases found in Portland cements.

As with Portland cements, hydration occurs by a process of dissolution and precipitation.This means that the anhydrous phases in the grains of cement dissolve to give ions insolution, much as sugar or salt dissolve in water. However, unlike sugar or salt, the ionsfrom the cement phases can recombine in different proportions with the ions of water.These new combinations come out of solution or precipitate, due to their lower solubility,as the hydrate phases.

When calcium aluminate cement (or monocalcium aluminate, CA) is placed in watercalcium ions (Ca2+) and aluminate ions (Al(OH) 4

– ) dissolve in the water to give a solution.These can combine as several different types of hydrate generally known by their chemicalformulas – CAH10, C2AH8, C3AH6 and AH3, plus poorly crystallized gel-like phases. Inorder to understand when and why the different hydrates form, and the practical consequencesof this, it is useful to recall some notions of thermodynamics.

When different phases are mixed together, they will tend to react and recombine togive a new mixture of phases that has the lowest energy – this is the stable phaseassemblage. However, sometimes it is quite difficult for the stable phases to form immediatelydue to the rearrangement of ions that needs to occur (nucleation). In this case it often

Calcium aluminate cements 2/7

happens that different phases form temporarily, which are metastable. They have a lowerenergy than the starting phase assemblage, which provides a driving force for theirformation, but they have a higher energy than the stable phase assemblage. Thus a drivingforce remains for the meta-stable phases to in turn react to give the stable phase assemblage.As illustrated in Figure 2.5, this may be thought of as a ‘step’ in energy. The appearanceof metastable phases is fairly common in other systems such as aluminium alloys of thetype used in aircraft.

Figure 2.5 Scheme indicating the meaning of metastable and stable hydrates.

Ene

rgy

Unstablestarting

state:anhydrous

phases

Stablefinalstate:stablehydrates

Metastableintermediatestate:metastablehydrates

At all temperatures the stable hydrates of CA are C3AH6 (a form of hydrogarnet) andAH3, (gibbsite):

3CA + 12H ⇒ C3AH6 + 2AH3 (2.1)

The crystal structure of C3AH6 is cubic so it usually has a morphology of compactequiaxed facetted crystals, (Figure 2.6), whereas AH3 often is poorly crystalline and isdeposited in formless masses. Figure 2.7 shows the microstructure of a 30-year-oldconcrete from a real structure (bridge at Frangey built 1969 with total w/c ~ 0.4). TheC3AH6 is finely dispersed in a matrix of hydrated alumina.

Figure 2.6 Morphology of C3AH6 crystal. The crystals in this picture have formed at a high water-to-cementratio, so are quite large. Those formed in normal pastes during conversion are typically less than a micron insize.

2/8 Calcium aluminate cements

As these are the stable phases, other phases that may form initially will over timetransform or convert to these phases.

At temperatures up to about 65°C the nucleation of C3AH6 is very slow. For temperaturesup to about 27–35°C the first hydrate to form is usually CAH10, while C2AH8 and AH3

dominate initially between 35 and 65°C (Fryda et al., 2001):

CA + 10H ⇒ CAH10 (2.2)

2CA + 16H ⇒ C2AH8 + AH3 (2.3)

The subsequent conversion reactions to the stable phases are:

2CAH10 ⇒ C2AH8 + AH3 + 9H (2.4)

3C2AH8 ⇒ 2C3AH6 + AH3 + 9H (2.5)

All these reactions take place through solution, that is, the reacting phases dissolveand the product phases are precipitated from solution. Although thermodynamically it ispossible for CAH10 to transform directly to C3AH6, C2AH8 will usually form as anintermediate phase due to ease of nucleation. Once the stable phases are present these willcontinue to form even if there is a fall in temperature.

From a practical point of view, the formation of these different hydrates is importantbecause they have very different densities and contain different amounts of combinedwater (Table 2.1). Thus when the conversion of metastable to stable hydrates occurs:

1 There is a reduction in solid volume and so an increase in porosity and reduction instrength at an equivalent degree of hydration (see section 2.5)

2 Water is released which is available to hydrate any remaining anhydrous phases

The most straightforward analytical method to determine the type of hydrates presentis differential thermal analysis. The metastable phases have decomposition peaks in thetemperature range 100–200°C, while the stable phases decompose in the range 250–350°C (Figure 2.8).

Figure 2.7 Microstructure of 30-year-old CAC concrete from structure (w/c ~0.4) – fully converted. A finedispersion of C3AH6 (light grey) in a darker mass of hydrated alumina. There are very few, small isolatedpores. The bright areas are unreacted iron rich anhydrous phases.

Pore

Clusters ofC3AH6crystals

Mass ofhydratedalumina

2μm

Calcium aluminate cements 2/9

2.2.2 Impact of conversion

The impact of conversion is greatest at high water-to-cement ratios (≥ 0.7). In this case,if the temperature is kept low, there is sufficient water and space available for nearly allthe reactive anhydrous phases to react to give meta-stable hydrates (CAH10 and C2AH8).Because these have a low solid density they fill most of the space originally occupied bywater, leaving low porosity (Figure 2.9(a)). When conversion occurs the formation of themore dense stable hydrates (C3AH6 and AH3) leads to a considerable decrease in solidvolume, an increase in porosity and a consequent decrease in strength (Figure 2.9(b)).The volumetric relationships are shown schematically in Figure 2.9(c).

When the w/c ratio is kept low (i.e. < 0.4), there is insufficient water and spaceavailable for all the cement to react to form metastable hydrates (Figure 2.10(a)). In thiscase the water released by conversion is available to react with more of the cement to givemore hydrates. The net reduction in solid volume and increase in porosity is lessened andthus dense, low porosity microstructures are obtained after conversion (Figure 2.10(b)).The consequent decrease in strength is therefore less marked. Figure 2.10(c) shows thevolumetric relationships in this case.

Although the water-to-cement ratio is the principal parameter controlling the propertiesof all concretes, this impact on microstructure before and after conversion emphasizes itsspecial importance for calcium aluminate cement concretes.

Table 2.1 Density and combined water for calcium aluminate hydrates

Phase Density (kg/m3) Combined water (%)

CAH10 1720 53C2AH8 1750 40C3AH6 2520 28AH3 2400 35

Figure 2.8 DTA traces of metastable (unconverted) and stable (converted) CAC pastes.

Converted

Unconverted

gel CAH10C2AH8 AH3 C3AH6

100 200 300 400Temperature (°C)

2/10 Calcium aluminate cements

Figure 2.9 Microstructures and hydration scheme of CAC concretes at high w/c (~0.7). When hydrated at20°C ((a), top) a fairly dense microstructure can be produced, but this will convert to the coarse openmicrostructure as shown in (b) (middle) for a concrete hydrated at 70°C. The volumetric relationships areshown schematically in (c) (bottom).

Metastablehydrationproducts

Relic ofcement grain

all CA reacted

Fine pores

(a)

Pores

Pores

Cement(unreactive phases)

Cement(unreactive phases)

Cement

Unconvertedhydrates

(c)

Water

Convertedhydrates

Relic ofcement grain

all CA reacted

Stablehydrationproducts

Coarse pores

(b)

Calcium aluminate cements 2/11

Cement(unreactive phases)

CementCement

Unconvertedhydrates

Pores Pores

Convertedhydrates

(c)

Water

Figure 2.10 Microstructures and hydration scheme of CAC concretes at low w/c (~0.4). When hydrated at20°C ((a), top) a dense microstructure is produced, after conversion the microstructure remains dense withthe further hydration of unreacted cement ((b) middle). The volumetric relationships are shownschematically in (c) (bottom).

Stablehydrationproducts

Fine pores

Cement grain,less CAremaining

(b)

Cement grain,substantial CAremaining

Metastablehydrationproducts

Fine pores

(a)

2/12 Calcium aluminate cements

2.3 Properties of fresh CAC concrete – setting,workability, heat evolution

2.3.1 Setting

The setting times of calcium aluminate cements are similar to those of Portland cement,namely around 3–5 hours. For instance, the British standard for CAC, of the ‘Fondu’ type,requires that the initial set is not less than 2 hours and the final set not more than 8 hours,when measured on pastes at standard consistency (~24–28% water) by the Vicat needlemethod. The reason for this relatively long induction period before the onset of initialsetting is the already mentioned delay in the nucleation of the hydrates.

The setting time of calcium aluminate phases becomes progressively longer withdecreasing lime-to-alumina ratio. C3A, the lime rich phase present in Portland cements,reacts rapidly with water, necessitating the addition of gypsum to Portland cement clinkerto prevent flash setting. At the other extreme CA2 reacts extremely slowly and CA6 isinert at ambient temperatures. The setting time of pure CA is typically about 18 hours, butthe shorter setting time seen in calcium aluminate cements is due to the disproportionateeffect of the small quantities of lime-rich C12A7 phase present. Because of this strongeffect of C12A7, variations in its amount can cause the setting time to change by hours andexcessive amounts can cause early stiffening. The amount of C12A7 is therefore tightlycontrolled during production to avoid this variability in setting time and other propertiesthat may also be affected as a consequence.

The setting time of CACs may also be affected unintentionally by contaminants orthey may be intentionally adjusted by the use of additives. A number of contaminants,such as sugars, most acids, glycerine etc., have a serious retarding effect. Seawater alsotends to retard setting. However, used intentionally, retarders such as citric acid or sodiumcitrate have beneficial effects on fluidity and working time.

The most common, effective accelerators for CACs are lithium salts (e.g. lithiumcarbonate) and these may be combined with retarding workability aids, such as sodiumcitrate, to improve workability while maintaining a normal setting time.

Accidental contamination with Portland cement will also accelerate the setting time ofCAC (flash setting may occur if contamination is severe). However, this effect may beused in a controlled way, in complex formulated products.

As different hydrates form at different temperatures (as described above) the effect oftemperature on setting time is more complex than in Portland cements. In pastes mixedin small quantities it is observed that the setting time increases with temperature up toabout 27°C, before decreasing again as the temperature is further increased. However, inconcretes mixed in normal quantities this effect is less pronounced due to a range ofeffects including self-heating and the heating due to the friction during mixing of theaggregates present in the concrete (Figure 2.11).

2.3.2 Workability

In the absence of admixtures the intrinsic ease of placing CAC concretes is similar to thatof Portland concretes at the same water-to-cement ratio. CAC concretes appear ‘dryer’,but the temptation to add water must be avoided as they flow well under vibration.

Calcium aluminate cements 2/13

Classic plasticizers such as ligno-sulphonate and sodium citrate work reasonably wellwith CACs, but there may be a significant retarding effect. Superplasticizers based onnaphthalene and melamine sulphonate are not very effective. Until recently, the best wayof obtaining reasonable workability at the recommended low w/c ratios was to use a highcement content, typically more than 400 kg/m3.

For the same reason it is important to have a well-graded aggregate, particularlysufficient fine sand, to minimize the volume between the aggregates that the cement pasteneeds to fill.

The modern generation of superplasticizers (e.g. PCP type) are highly effective withCACs, but the dosage must be carefully controlled to avoid excessive retardation. At afree w/c of around 0.35 it is now possible to obtain flowing concrete for more than2 hours. These concretes still have strengths of around 30 MPa at 8 hours even at ambienttemperatures of only 5°C (Fryda and Scrivener, 2002).

2.3.3 Rate of reaction and heat evolution

At the end of the induction period rapid reaction occurs. In Portland cement the mainhydration product is deposited around the hydrating cement grains, so that after the firstfew hours the rate of reaction is progressively slowed down. In CACs, in contrast, thehydration products are deposited throughout space (Figure 2.12). The rate of reactiononly slows due to one of the reactants (water or cement) being used up or to lack of spacefor the hydrates to precipitate. This pattern of reaction leads to the rapid hardening ofCAC concrete and it is quite easy to obtain strengths of more than 20 MPa at 6 hours andover 40 MPa at 24 hours. The other consequence of the rapid reaction is that the heat ofhydration is evolved over a relatively short time. As the heat generation is rapid relativeto the rate of heat dissipation, self-heating may be considerable in sections more thanabout 100 mm in size. It is quite usual to have maximum temperature in the range 50–80°C or even higher. Figure 2.13 shows the temperature profile in a beam of concrete.Such self-heating effects are very important for the strength development on CAC concrete,as discussed below (section 2.5).

12

10

8

6

4

2

0

Initi

al s

ettin

g tim

e (h

)

Concrete

Paste

0 5 10 15 20 25 30 35 40

Temperature (°C)

Figure 2.11 Effect of temperature on setting times. From Cottin and George (1982).

2/14 Calcium aluminate cements

Despite these high temperature rises, CAC concretes do not appear to be overly susceptibleto thermal cracking. This may be due to the fact that creep is facilitated by the conversionreaction and relaxes thermally induced strains.

An important consequence of the rapid temperature rise is the need to keep the concretewet to prevent the drying and dehydration of the surface. For large sections it is recommendedto remove the formwork as early as possible (around 6 hours) and spray the surfacewith water.

One advantage of this self-heating is that concreting can be continued in periods ofvery cold weather and indeed at sub-zero temperatures, provided the concrete is notallowed to freeze before hydration commences.

2.3.4 Shrinkage

Intrinsically the shrinkage of calcium aluminate cement is roughly similar to that ofPortland cement at the same degree of hydration. However, as hydration occurs muchmore rapidly this shrinkage occurs over much shorter period of time and can lead toproblems with early age cracking (Figure 2.14).

In order to minimize these effects the same rules of concrete technology that areknown for Portland cements apply, namely:

Portland

Reacting grains

CAC

Hydration products

Figure 2.12 Schematic representation of the deposition of hydration products for the hydration of Portlandand calcium aluminate cement.

0 2 4 6 8 10 12 14 16 18 20 22

Hours after gauging

80

70

60

50

40

30

20

10

Tem

pera

ture

(°C

)

Figure 2.13 Temperature rise in beam 35 × 180 cm of CAC concrete. Formwork was removed after 5 hrsthen beam sprayed with water for 20 hours. Adapted from French Standard (1991).

Calcium aluminate cements 2/15

• prolonged moist curing• good aggregate grading to minimize paste volume• limitation of bay sizes (to around 3 m in unreinforced slabs)• use of polymeric fibres in thin sections.

2.4 Strength development

Due to its hydration, with the successive formation of metastable and stable hydrates, thestrength development of CAC concretes is more complicated than that of Portland cementconcrete. However, there are a few straightforward facts to remember:

1 Conversion is a thermodynamically inevitable process, so the converted strength shouldbe used for design.

2 Once conversion has occurred, stable hydrates are present and therefore the strength isalso stable and is likely to increase due to further hydration of any remaining unreactedcement.

3 As with Portland cement concretes, and indeed any other porous material, the majorfactor related to strength is the porosity (or lack of it) and hence the water/cement ratioand the degree of reaction.

4 At equivalent water to cement ratios the converted strength of CAC concrete is slightlylower than that of standard Portland cement concretes.

For small elements maintained at temperaures around 20°C or lower, strength developmentfollows the pattern shown in curves A and B in Figure 2.15 – this is the pattern of strengthdevelopment often associated with the conversion of calcium aluminate cements showingsubstantial loss of strength. In fact it is rare to find such extreme cases in practice due tothe substantial self-heating effects already discussed. Even in such extreme cases it isimportant to realize that the strength decreases to a stable minimum depending on thewater-to-cement ratio and does not continue decreasing to zero.

In such cases the initial hydration reactions during setting and hardening produce a

0 7 14 21 28

800

700

600

500

400

300

200

100

0

Shr

inka

ge (

μm/m

)

CAC

PC

Age (days)

Figure 2.14 Comparison of shrinkage for CAC and OPC concrete measured in 40 × 40 × 160 prisms at20°C, 50% RH. Data from French Standard (1991).

2/16 Calcium aluminate cements

microstructure of metastable hydrates that combine a large amount of water and have arelatively high specific volume (low density). They form throughout space (as explainedpreviously) and therefore create a microstructure wherein the porosity is very low (Figure2.9(a)). Consequently, it is possible to obtain quite high strengths even at high water-to-cement ratios.

However, as the metastable hydrates convert to the stable higher density phases withless water in their structure there is a decrease in volume of hydrates and so an increasein porosity and a decrease in strength. For concrete maintained at ambient temperaturesof 10–20°C this process may take ten years or more (still a comparatively short time inthe life of most concrete).

Once conversion has occurred the strength is stable, but if the original water-to-cementratio is low, the strength may increase after the minimum as water released by hydrationreacts with remaining unreacted cement.

In practice, for concrete sections more than a few tens of centimetres in size, the heatgenerated by hydration will cause the temperature of the concrete to rise, typically to 50–80°C or more. In these conditions the stable hydrates are quickly formed since highertemperatures accelerate the rate of conversion as shown in Figure 2.16. For isothermalcuring at 38°C the minimum strength occurs at 5 days. At 50°C the minimum is reachedby around 24 hours and for temperatures greater than 50°C conversion occurs within afew hours.

In the case of many civil engineering concretes the size of the pour is sufficient forstable hydrates to be formed during the initial phase of curing and consequently there isno regression in strength, which will continuously increase with increasing hydration.This situation is illustrated in curves C and D of Figure 2.15.

2.4.1 Importance of water-to-cement ratio and control ofconcrete quality

As previously stressed, the control of water to cement (w/c) ratio is particularly importantin CAC concretes due to its impact on the stable converted strength. The converted

D – w/c ~0.8, with self-heating

Time

B – w/c ~0.7, ~20°C,no self-heating

C – w/c ~0.4, with self-heating

Portland w/c ~0.4A – w/c ~0.4, ~20°C,no self-heating

Str

engt

h

Hours Days Months Years

Figure 2.15 Schematic strength development of CAC concrete at high and low w/c, under isothermalcuring at 20°C and under conditions of self-heating, compared with Portland cement concrete.

Calcium aluminate cements 2/17

strength of CAC concrete decreases as w/c increases (Table 2.2). Historically total ratherthan free w/c ratios were used in studies of CAC concrete. Approximate equivalent freew/c are given in the table (assuming an aggregate absorption of around 1.2%). However,as CAC reacts much faster than Portland cement, there is less time for water to beabsorbed by the aggregates and the effective w/c may be higher.

Figure 2.16 Impact of temperature of rate of conversion and strength development for isothermal curing.Adapted from French et al. (1971).

Table 2.2 Variation in converted strength with water-to-cementratio (from George, 1976)

Total w/c Approx free w/c Minimum convertedstrength

0.33 0.28 530.4 0.35 470.5 0.44 370.67 0.58 27

Based upon considerable experience and laboratory testing, recommendations for ‘totalw/c’ ratios were established. These manufacturer and industry recommendations are thatthe maximum total w/c ratio should be no greater than 0.4 (equivalent to free w/c ratio of~ 0.33–0.36). With this maximum w/c the converted CAC concrete has good residualstrength and low permeability.

However, the consequences on stable strength of using a high w/c ratio may be maskedif the concrete strength is controlled at early ages on small samples cured at ambienttemperature – because such a concrete would contain metastable hydrates. These samplesmay exhibit relatively high strengths, but this is misleading since they will eventuallyundergo a significant strength loss with time due to conversion.

120

100

80

60

40

20

0

Str

engt

h (M

Pa)

80°C

50°C

30°C

25°C

38°C

25°C30°C38°C50°C80°C

0.001 0.01 0.1 1 10 100 1000Age (days)

2/18 Calcium aluminate cements

The control of the strength of calcium aluminate cement as delivered on-site shouldnot be made by casting cubes or cylinder in metal or cardboard moulds. For the reasonsalready discussed, control specimens cured at or near 20°C will have high unconvertedstrengths after one day, which are not representative of the ultimate strength of the in-situconcrete. Being insensitive to the high water-to-cement ratios used, such samples wouldgive a false representation of the long-term strength.

A well-proven procedure that gives an accurate indication of the long-term designstrength is to place control specimens immediately at a temperature of 38°C at 100%R.H. (or sealed). The minimum, converted, stable strength will occur at 5 days.

Alternatively insulated moulds can be used, in which the heat of hydration is retained,which means that the temperature profile of the control specimens is similar to that of theconcrete in situ. Such test methods can give a useful forecast of the minimum strength butshould not be relied upon without reliable quality assurance of the water addition for allbatches on-site.

Site-verification acceptance of batched concrete is an issue for most concrete irrespectiveof cement type. This is most often undertaken by a combination of batching/mixing QAprocedures and acceptance testing. For CAC concretes, the assurance of conformity withthe maximum w/c is particularly important for long-term performance. Where CACconcrete is being used to shorten the time before reintroducing critical elements intoservice, the time required for conventional physical testing may be inappropriate andother routes may need to be considered. However, CAC concrete can be used with confidenceif appropriate assurance of conformity can be properly achieved.

2.4.2 Other factors affecting conversion and strengthdevelopment

As conversion is a chemical reaction occurring through solution, the presence of moistureis essential for it to occur. However, it is unlikely and unreasonable to assume that anyconcrete could be maintained in sufficiently dry conditions for conversion to be avoided.This is illustrated in Figure 2.17, which shows little difference in the strengths of concrete

Air/water curedAir curedWater cured

AggregateLimestoneSiliceous

3 months 6 months 3 months 6 months

80

70

60

50

40

30

20

10

0

f’ c (

MP

a)

Figure 2.17 Impact of storage conditions on compressive strength of CAC concrete cured at 40°C.From Lamour et al. (2001).

Calcium aluminate cements 2/19

at w/c = 0.4 stored in either air or water at 40°C. Subsequent storage of the air curedspecimens in water, led to an increase in strength, probably through hydration of remaininganhydrous material.

In recent years several proposals have been made for systems in which conversion isinhibited (e.g. Majumdar and Singh, 1992; Osborne, 1994; Ding et al., 1996). These relyon additions of material containing reactive silica – slag, silica fume, metakaolin, etc. Fortemperatures below about 50°C, the stable phase, strätlingite (C2ASH8), forms instead ofC3AH6 by the reaction of the metastable hydrates with silica and, if the amount ofaddition is high enough, little or no strength regression due to conversion occurs. However,very high additions of these materials impair the early strength development. Also, if theconcrete experiences temperatures above 50°C, as may occur through self-heating, C3AH6

will still be formed.Another factor with an important impact on strength is the kind of aggregate used.

Considerably higher strengths are obtained with calcareous aggregates, compared tosilicons aggregates.

2.5 Other engineering properties

There are relatively few studies of other engineering properties at w/c ratios around therecommended 0.4. A recent study showed that the tensile strength and toughness wereroughly the same as for Portland cement concretes (of comparable compressive strength)(Lamour et al., 2001) and that these properties tended to be slightly less affected byconversion than the compressive strength. As for Portland concrete, the aggregate has amajor impact on the elastic modulus, which otherwise is similar to Portland cementconcretes of a similar strength. Poisson’s ratio varied between 0.15 and 0.25, againsimilar to Portland concretes.

Relatively few studies of creep are reported, especially in the w/c range now used.However, it appears that the rate of creep depends on the applied stress/strength ratio ina manner similar to that of Portland cement.

2.6 Supplementary cementing materials

A major trend in Portland concrete technology is the increasing use of supplementarycementing materials – fly ash, slag, silica fume, fine limestone, metakaolin, etc. – toimprove durability and reduce clinker consumption. As mentioned above, the presence ofreactive silica favours the formation of strätlingite. In moderate amounts this may have apositive effect on the minimum, stable strength although the high additions necessary tocompletely eliminate strength regression lead to poor early strengths.

In other regards the impact of such additions on calcium aluminate cement has notbeen extensively studied. There is some evidence that they may contribute to improvementsin rheology and allow the formulation of self-placing concretes.

2.7 Durability/resistance to degradation

As with all concretes, the primary factor determining durability is a good-quality concretewith low porosity. This implies the use of low water-to-cement ratios. Unconverted CAC

2/20 Calcium aluminate cements

concretes have exceptionally low permeability. Although the permeability increases withconversion reasonably low levels can still be achieved in converted concrete made withtotal w/c ratios at or below 0.4. Studies of old structures indicate that CAC concrete formsa dense surface layer, which may contribute to the good performance for over 70 yearsseen in these massive structures even with total water-to-cement ratios around 0.6 (Scriveneret al., 1997a).

In addition to the specific forms of degradation discussed below, it should also benoted that no cases of alkali aggregate reaction have ever been observed with CACconcrete, due to its lower alkalinity (around pH 12.4 in comparison to >13 for Portland;Mácias et al., 1996).

2.7.1 Reinforcement corrosion

Corrosion of reinforcement is well known to be the major cause of durability problems inPortland cement concrete. This is due to impairment of the passive oxide film, whichnormally protects reinforcing steel in the alkaline environment of the concrete, due toeither reduction in this alkalinity through carbonation or the penetration of chloride ions.

Calcium aluminate cement concrete has a slightly lower alkalinity than Portland concrete,but the pH is still more than high enough to ensure the formation of a passive film on steeland so protection against corrosion under normal circumstances. The equilibrium pH forthe metastable phase assemblage is 12.13 while that for the stable phase assemblage isonly slightly different at 11.97 (Damidot, 1998). In real concretes the pH levels are higher(~12.5), due to the small amounts of alkali metal oxides impurities present, which remainin the pore solution (Mácias et al., 1996). Conversion does not therefore change theintrinsic capacity of CAC to protect reinforcement.

As with Portland cement, both carbonation and penetration of chloride ions may leadto the breakdown of the passive layer, and to active corrosion occurring. Again, as withPC concrete, these processes are controlled by transport of CO2 or chloride through theconcrete cover, the rate of which is controlled by the resistance to permeability or diffusionof the concrete cover, which is in turn primarily a function of its porosity and hence thew/c ratio used and the degree of curing. The conversion in CAC concretes then becomesimportant in that it can lead to a highly porous concrete if a high w/c ratio is used.

Field evidence demonstrates that CAC concrete is capable of providing good protectionto reinforcement over a sustained period of time. One of the largest structures ever builtwith CAC concrete was part of the ocean harbour in Halifax, Nova Scotia, Canada. Some6000 tonnes of CAC was used for its construction between 1930 and 1932. Cores takenfrom the structure in the early 1990s showed that the reinforcing steel was in goodcondition and that penetration of chloride and sulfate was generally low. The formationof a dense surface layer at the surface of the concrete is thought to have contributed to itsgood durability.

2.7.2 Sulfate attack

CACs were originally developed for improved sulfate resistance and early CAC concretewas successfully used in tunnelling works through gypsum deposits. CAC concretes have

Calcium aluminate cements 2/21

consistently performed extremely well in the field, both in structures and in trials, providedthe water-to-cement ratio is kept to the recommended levels. More variable results havebeen obtained in laboratory studies, but in such studies the pH of the storage solution isoften not adjusted to keep it at the relatively low levels found in the field. Removing thesurface layer of cubes also had a detrimental effect on their resistance to sulfate solutions.For more detail see Scrivener and Capmas (1998).

2.7.3 Freeze–thaw damage

Little laboratory data exists on the freeze–thaw resistance of calcium aluminate cement.Those studies that have been made indicate that CAC concretes show good freeze–thawresistance when the porosity is low (concretes with a total w/c ratio less than 0.4). Fieldperformance, especially where CAC concrete has been used in cold climates, have notidentified any particular problem of freeze–thaw resistance. CAC concretes can be airentrained.

Studies of CAC concretes made with special calcium aluminate aggregate (AlagTM)have shown that this type of concrete has exceptionally good freeze–thaw resistance. Avery particular application where very good performance is achieved is in areas wherethere is spillage of liquid gasses (T ~ –170°C, Figure 2.1).

2.7.4 Alkaline hydrolysis

A form of degradation that is particular to calcium aluminate cements is alkaline hydrolysis.This is a very rare form of degradation that only occurs in concretes of very high w/c(> 0.8). In addition, an external source of alkalis and moist conditions are required. Themechanism is believed to be the leaching of alumina in highly alkaline conditions. Manyexplanations also focus on the role of carbonation. The role of CO2 in exacerbatingdegradation by alkalis is probably that carbonation leads to a breakdown of the calciumaluminate hydrates into calcium carbonate and alumina, thus making the alumina moresusceptible to alkaline leaching in highly porous concretes.

Microstructural examination reveals that concrete suffering from this form of degradationis extremely porous, leading to dramatic strength loss. The source of alkali can be fromthe aggregates or external, for example water percolating through Portland concrete orfibre boards or alkaline cleaning fluids. Laboratory studies followed for nearly 30 yearsconfirm the critical role of w/c. In two series of specimens – one with alkali-containingaggregate and the other where water was absorbed through Portland concrete – degradationonly occurred for the highest water-to-cement ratios – 1.0 and 1.2.

Long-term field studies have shown no problems in situations where CAC concrete (ofnormal w/c) adjoins Portland concrete (Deloye et al., 1996).

2.8 Structural collapses associated with CAC concrete

Many civil engineers will have heard of the structural collapses involving CAC concretethat occurred in the UK in the 1970s. These are treated in detail in a recent Concrete

2/22 Calcium aluminate cements

Society (1997) report. The origin of these events was the use of CAC concrete in pre-cast,pre-stressed beams, which occurred in the post-war construction boom. At the time,conversion was known about, but the advice was conflicting. Rather than conversionbeing regarded as inevitable, it was considered to be negligibly slow if warm moistenvironments were avoided (Code of Practice 114 (1957) and annex (1965).

The importance of low water-to-cement ratios was also known, but recommendations(Code of Practice annex 114, 1965) to keep the w/c to 0.5 or less conflicted with therequirement of a cement content less than 400 kg/m3, to limit the effects of self-heating.Concrete containing less than 400 kg/m3 of cement would be most difficult to place atw/c < 0.5.

Nevertheless, loss of strength due to conversion was not the sole or primary cause offailure. In two out of the three incidents design problems were the primary cause – lackof adequate bearing area between the beams and the wall supports, which may have beenexacerbated by thermal movement of the beams, which were not adequately linked to thewalls. In one of these cases the original beams were tested to destruction and found tohave adequate strength for the design, the remaining beams being re-used in the rebuilding.In the third case, the concrete was found to have degraded due to external sulfate attackfrom adjacent plaster. This attack was facilitated by the high porosity of the concrete dueto the use of a high w/c.

In the wake of these events, occurring shortly after a major structural collapse of atower block built with OPC concrete (Ronan Point), measures were taken which effectivelyprevented the use of CAC concrete in structural applications. The overall number ofbuildings containing CAC concrete built during this period which continue in service isbelieved to be around 30 000–50 000. After the collapses some 1022 of these were inspected,of which 38 were identified as having problems, but in only one of these was thisattributed to loss of strength due to conversion.

Pre-cast, pre-stressed CAC concrete beams from a similar era were also implicated inproblems in Spanish apartment buildings in the 1990s. The worst of these was the collapseof the ‘Tora de la Peira’ apartment block, in which beams made with both CAC andPortland concrete failed (George and Montgomery, 1992). In these cases the CAC concretewas degraded by alkaline hydrolysis due to poor-quality aggregates and highly porousconcrete.

These unfortunate episodes illustrate the importance of understanding the materialsbehaviour of CAC concrete. The use of CAC concrete in small sections, where conversiondoes not occur rapidly due to self-heating, raises the difficulty of ensuring that the water-to-cement ratio is strictly controlled. Nowadays, steam-curing techniques mean that pre-cast beams with high early strength can be satisfactorily produced in Portland concrete atmuch lower cost than with CAC. It is unlikely that primary ‘bulk’ structural use of CACconcrete would be viable or justified.

The conclusion of the Concrete Society working party on CAC in construction whichreported in 1997 (Concrete Society, 1997) was that there were situations in which theCAC concrete could be used safely and advantageously in construction, provided that itslong-term converted strength is taken into account for design purposes. This has led to achange in the guideline to the UK building regulations, allowing use of materials whoseproperties change over time, provided that their residual properties are predictable andadequate for the design. However, there are still no BS codes of practice for the structuraluse of CAC concrete.

Calcium aluminate cements 2/23

2.9 Modern uses of CAC concrete

The modern uses of CACs are based on areas where their special properties provide amajor advantage over other cements.

2.9.1 Resistance to acids/sewage networks

The excellent resistance to acids, particularly those produced by bacteria, leads to use insewage networks (Figure 2.1) (Scrivener et al., 1999; Letourneux and Scrivener, 1999).In sewers, bacteria in the effluent produce hydrogen sulphide gas (the smell of rotteneggs), which is in turn used as a nutrient by another group of bacteria, resulting in theproduction of sulphuric acid. This can lead to very severe degradation of concrete, particularlyin the crown and along the water-line of pipes. The growth of bacteria is most favourableat 30°C so such attack has become a notable problem in countries with hot climates.

The extensive positive field experience with CAC pipes is supported by the results ofa 12-year field trial in South Africa (Goyns, 2001). The images shown in Figure 2.18clearly illustrate the superior performance of CAC concrete. Reinforced pipes of 60 mmof different materials were placed along the same length of sewer. After 12 years thePortland concrete with siliceous aggregate has been completely eroded through the 60 mmthickness so that the earth is visible (Figure 2.18(a)). A section made with Portlandcement and dolomitic limestone showed only marginally better performance.

In contrast, the section made with CAC concrete (with siliceous aggregate) was in amuch better condition (Figure 2.18(b)) with a maximum depth of attack of around 10–15 mm only at the water-line. The reasons for this improved performance are:

(a) (b)

Figure 2.18 Sewage pipes after 12 years in a field trial in South Africa. The Portland/siliceous aggregatesection (a) has completely eroded through its 60 mm thickness in places. The CAC/siliceous aggregatesection (b) shows maximum erosion of 10–15 mm only at the water-line.

2/24 Calcium aluminate cements

• The presence of hydrated alumina, which is insoluble in acid down to pH 4 and has ahigh neutralization capacity at lower pHs

• The inhibition of bacterial activity, which limits the surface pH to around pH 3.5 incontrast to pHs around 1 which develop on Portland concrete.

Studies have shown that even better resistance can be provided by using CAC in conjunctionwith a sacrificial dolomitic aggregate or (best of all) a synthetic aggregate (AlagTM).

2.9.2 Resistance to abrasion and impact

CAC concretes show very good resistance to abrasion and impact when made with suitablehard aggregates and particularly with AlagTM, which is a synthetic aggregate made fromCAC clinker. This property has led to their use in a number of applications including orepasses in South Africa, Canada and Australia; in hydraulic dams in Switzerland, Franceand Peru and in sections of roads subject to heavy wear such as motorway toll booths(Scrivener et al., 1999).

For Portland concrete it is generally found that abrasion resistance is a simply relatedto compressive strength. CAC concrete shows superior resistance to abrasion comparedto Portland concretes of similar strength (Figure 2.19) (Saucier et al., 2001). This

NormalCAC/Alag – 69

MPa

VHS CAC/Alag– 133 MPa

VHS PC conc.– 135 MPa

PC+SF conc.– 50 MPa

Normal PC conc.–20–30 MPa

7

6

5

4

3

2

1

0

Rel

ativ

e ab

rasi

on r

esis

tanc

e (g

lass

= 1

)

Figure 2.19 Relative abrasion resistance of CAC/Alag and Portland cement concretes. The very highstrength (VHS) concretes have similar strength but very different abrasion resistances. Data from Saucieret al. (2001).

improvement is believed to be due to the improvement in the quality of the interfacialtransition zone (ITZ) between the cement paste and aggregate. Normally in Portlandconcretes the ITZ is more porous, due to the difficulty of packing the cement grains inthis region (‘wall effect’); as the CAC hydrates form throughout space they infill thisregion better as illustrated in Figure 2.20. Partial reaction at the surface of the syntheticaggregate also improves the bond.

Calcium aluminate cements 2/25

2.9.3 Rapid strength development

CAC concrete can provide unique technical advantages, when very rapid strengthdevelopment is needed, for example industrial applications where a rapid return to serviceis important. Another example is the pre-cut tunnelling method where a CAC/slag blendis used to provide a temporary concrete lining, prior to excavation of the main tunnel bore(Defargues and Newton, 2000). The concrete is applied as shotcrete and develops astrength of over 8 MPa in 4 hours, which allows the next tunnel segment to be excavated.A lining of conventional concrete is installed in the final tunnel.

The rapid strength development of CAC concrete is particularly advantageous in coldclimates and it has been widely used in the Arctic – for example, for grouting cable stays.Even at 0°C CAC concrete can develop strengths of 20–30 MPa at 16 hours (Scrivenerand Capmas, 1998).

2.9.4 Thermal resistance

CAC concrete is very resistant to heating and thermal shock. This property is explainedin more detail in a later chapter, which discusses heat-resisting and refractory concretes.However, this property is also used to provide resistant floors in (for example) foundries(Figure 2.1). Another use drawing on this property is in buildings for fire training, whichcan withstand repeated lighting and extinguishing of fires (Figure 2.1).

Figure 2.20 Interfacial transition zone (ITZ) between CAC and Alag aggregate. Due to formation ofhydrates throughout space the ITZ has low porosity, unlike in Portland cement concretes.

2/26 Calcium aluminate cements

2.10 Use of CACs in mixed binder systems

An increasingly important use of calcium aluminate cements today is as a component ofmixed binder systems for specialist applications. The ingredients are usually blended dryand sold as mortars ready to mix with water. The applications are typically non-structuralfinishing operations in buildings, such as floor levellers, tile adhesives or fixing mortars.Another application is rapid-hardening repair mortars.

Such mixed binders are usually based on combinations of three reacting ingredients –calcium aluminate cements, Portland cement and calcium sulfate. The calcium sulfatemay be in the form of gypsum, anhydrite or plaster (calcium sulfate hemi-hydrate).Fillers such as limestone and supplementary cementitious materials such as slag may alsobe present. In addition these mixes usually contain several admixtures in dry form(accelerators, retarders, fluidifiers, plasticizers, thickeners). Fine siliceous sands (andsometimes fine aggregate) are also normally used in these special mortars as they wouldbe in conventional mortars.

Figure 2.21 shows the common zones of formulation of the CAC, PC and C$ components.Zone 1 concerns blends based on binary mixtures of Portland and calcium aluminatecement (with more of the former) to which smaller amounts of calcium sulfate may beadded. These mixtures are rapid setting and hardening.

C$Gypsumanhydrite

plaster

Zone 2

Zone 1

CAC PCRelative weights

Figure 2.21 Zones of formulation in ‘Building Chemistry’.

It has long been known that additions of CAC to Portland cement can induce flashsetting. The basis of this behaviour is that the added calcium aluminates combine with thecalcium sulfate in the Portland cement added during grinding. This disrupts the normalset control process of the Portland cement, in which calcium sulfate reacts with therapidly reacting C3A phase of the Portland clinker, producing products which coat thesurface of the C3A and block its further reaction for several hours. The relative rates ofdissolution of the various forms of calcium sulfate (anhydrite, hemi-hydrate and gypsum)are different, so the amount of CAC that needs to be added to give flash setting varieswith the amount and type of calcium sulfate in the Portland cement.

Such simple binary mixtures may show rather poor long-term strength development.This can be improved by the addition of admixtures or gypsum. Further discussion of themechanisms occurring in such system can be found in Amathieu et al. (2001). As anexample a correctly formulated mixture in this zone can have a working time of 2 hoursand shows strength of around 20 MPa after 6 hours.

Calcium aluminate cements 2/27

In such systems the hydrated microstructure is similar to that of Portland cement, withcalcium silicate hydrate being the dominant phase. The quantity of calcium hydroxidewill be decreased and the quantities of ettringite and/or monosulfate increased accordingto the relative amounts of CAC and calcium sulfate contained in the blend. For theamounts of CAC typically used, the metastable calcium aluminate hydrates are not formedso there is no question of conversion.

The mixtures in Zone 2 are composed of roughly equi-molar amounts of monocalciumaluminate (CA) from CAC and calcium sulfate, with minor amounts of Portland cement.The predominant hydration reaction in this zone is:

3CA + 3C$ · Hx + (38-x)H ⇒ C3A · (C$)3 · 32H + 2AH3 (2.6)ettringite

In equation (2.6) the formula AH3 for gibbsite is used for convenience, but this phase isoften poorly crystalline. Microstructural examination shows crystals of ettringite dispersedin a dense alumina-rich matrix. Both these phases are stable phases and will not spontaneouslyconvert to other phases.

Such mixtures can be controlled to be fluid for a few hours to allow placing and thenharden rapidly. However the main difference, between the mixtures of this zone and thoseof zone 1 is their capacity to combine water in the hydrates and so to rapidly reduce therelative humidity. This leads to their major application as floor-levelling compounds(Figure 2.22), where the rapid reduction of relative humidity (commonly referred to asdrying) allows them to be overlaid with carpet or floor tiles with a minimum of delay (aslittle as 8 hours). Table 2.3 shows the percentage of combined water by weight in varioushydrate phases.

Figure 2.22 Application of floor-levelling compound.

2/28 Calcium aluminate cements

2.10.1 Stability of ettringite-containing systems

As discussed, ettringite is an important component in all these mixed systems and questionsare posed as to its stability, regarding temperature, humidity, and carbonation. On the firsttwo issues, recent work published by Zhou and Glasser (2001) confirms that ettringite hasbroadly similar stability to temperature and humidity as C–S–H the principal hydratefound in hydrated Portland cement. It is important to realize that ettringite occurs as partof a system of hydrates, existing as a rigid, if porous, solid and not as an isolated powder.Thermogravimetric analysis shows that the decomposition temperatures of ettringite andC–S–H are very similar, around 110–130°C.

On the question of carbonation, analogies may also be drawn with conventional Portlandsystems. Finely powdered ettringite will carbonate rapidly, as will finely powdered C–S–H, but when either are part of a hydrate system the kinetics of the reaction are controlledby the ingress of CO2 into a porous solid. The carbonation of ettringite eventually leadsto calcite, gypsum and alumina:

3CaO · Al2O3 · 3CaSO4 · 32H2O + 3CO2 ⇒ 3CaCO3 + 3CaSO4 ·2H2O (2.7)

+ 2Al(OH)3 + 9H2O

This reaction leads to the release of water, which helps to maintain a high relativehumidity inside the solid, which impedes the ingress of CO2 gas. However, the loss ofwater corresponds to a decrease in the solid volume and an increase in porosity. Therefore,there will be some loss of strength in materials containing ettringite that become completelycarbonated. For formulations in Zone 1, such strength losses are negligible and will becompensated for by the continuing increase in strength due to hydration. For formulationsin zone 2 the loss in strength is more significant, but not catastrophic. Laboratory studies(Scrivener et al., 1997b) of accelerated carbonation of a composition which had a one-day strength of ~30 MPa showed that the samples stored at 100% CO2 and 60% R.H. hadstable strength of 35 MPa at 3 months (corresponding to complete carbonation) as opposed toa strength of 45 MPa for the reference sample stored at 60% R.H. with atmospheric CO2.

2.10.2 Dimensional stability of ettringite-containing systemsand shrinkage compensation

Another issue related to ettringite is that of dimensional stability. By formulation it is alsopossible to engineer the initial hydration reaction to produce controlled expansion and so

Table 2.3 Percentage of water combined in various hydrates

Phase Formula % H2O

ettringite 3CaO · Al2O3 · 3CaSO4 · 32H2O 46CAH10 CaO · Al2O3 · 10H2O 53monosulfate 3CAO · Al2O3 · CaSO4 · 12H2O 35gibbsite Al(OH)3 35C2AH8 2CaO · Al2O3 · 8H2O 40C3AH6 3CaO · Al2O3 · 6H2O 28CSH 1.7CaO · SiO2 · 4H2O 30Portlandite Ca(OH)2 24

Calcium aluminate cements 2/29

compensate for the normally observed drying shrinkage of Portland cement. This issimilar to the actions of shrinkage compensation in type K cements.

It is well known that the formation of ettringite is also related to undesirable expansionduring external sulfate attack and in so-called internal sulfate attack or delayed ettringiteformation (DEF) which may occur after exposure to temperatures >70°C during curing.However, the relationship between ettringite formation and expansion is extremely complex.For example, after exposure to temperatures >70°C during curing delayed ettringite formsin Portland cement mortars or concretes, irrespective of whether expansion occurs or not.Extensive studies have shown that the crystallization of ettringite only leads to expansiveforces if it forms in small pores (~50 nm) (Taylor et al., 2001).

Generally it can be said that the expansive formation of ettringite requires:

1 A source of alumina (e.g. calcium alumino monosulfate) distributed in small pores2 A source of sulfate, which can locally produce conditions that are supersaturated with

respect to ettringite formation.

In the formulated systems described above, such conditions may arise if there is an excessof calcium sulfate present, such that this does not all react to form ettringite during theinitial reactions. Systems that are correctly formulated have been shown to be dimensionallystable over several years, even if subject to wetting and drying cycles.

2.11 Summary

The aim of this chapter has been to introduce the reader to the properties and applicationsof calcium aluminate cements. Such cements have now been produced for nearly a centuryand are the most important type of special cement in terms both of the quantity used andtheir range of applications. Due to their higher price, they should not be considered as adirect replacement for Portland cements. Today the major applications of these cementsare in refractory concretes (described in a separate chapter) and as a component of pre-mixed binder systems used in the building industry, with lesser amounts being used inconventional concrete form.

In conventional concrete form it is important to take account of the phenomenon ofconversion whereby the concrete attains its stable long-term strength. If the concrete iskept cool, this process may take several years, but when self-heating occurs duringhardening, it is relatively rapid. It is important to maintain strict control on the water-to-cement ratio used to ensure adequate strength after conversion – it is recommended thatthe total w/c ratio should be at or below 0.4.

The principal properties of interest of CAC concrete are good resistance to chemicaland mechanical attack and rapid strength development.

As a component in mixed binders, the presence of CAC contributes to the formationof ettringite, which can be controlled to give systems with rapid hardening, shrinkagecontrol and rapid humidity reduction (‘drying’).

2/30 Calcium aluminate cements

References

Amathieu, L., Bier, T.A. and Scrivener, K.L. (2001) Mechanisms of set acceleration of Portlandcement. In Mangabhai, R.J. and Glasser, F.P. (eds), Calcium Aluminate Cements 2001. IOMCommunications, London, pp. 303–317.

Code of Practice 114 (1957) and annex (1965) The Structural Use of Reinforced Concrete inBuildings. British Standards Institution, London.

Concrete Society (1997) Calcium Aluminate Cement in Construction: A Re-assessment. TechnicalReport No. 46, Concrete Society, Slough.

Cottin, B. and George, C.M. (1982) Reactivity of industrial aluminous cements: an analysis of theeffect of curing conditions on strength development. Proceedings of the International Seminaron Calcium Aluminates, Turin 1982, pp. 167–70.

Damidot, D. (1998) Private communication.Defargues, D. and Newton, T. (2000) The Ramsgate road tunnel. Concrete April, 43–45.Deloye, F.-X., Lorang, B., Montgomery, R., Modercin, I. and Reymond, A. (1996) Comportement

à long terme d’une liaison Portland-Fondu. Bulletin des Laboratoires des Ponts et Chaussées–202 (March/April), 51–59.

Ding, J., Fu, Y. and Beaudoin, J.J. (1996) Effect of different organic salts on conversion preventionin high-alumina cement products. J. Adv. Cement Based Materials 4, 43–47.

French Standard (1991) Use of Fondu aluminous cements in concrete structures (annex). AFNORP15–316.

Fryda, H., Scrivener, K.L., Chanvillard, G. and Féron, C. (2001) Relevance of laboratory tests tofield applications of calcium aluminate cement concretes. In Mangabhai, R.J. and Glasser, F.P.(eds), Calcium Aluminate Cements 2001. IOM Communications, London, pp. 227–246.

Fryda, H. and Scrivener, K.L. (2002) User friendly calcium aluminate cement concretes for specialistapplications. In Proceedings International Conference on the Performance of ConstructionMaterials in the New Millennium, Cairo, Eygpt, February.

Füllmann, T., Walenta, G., Bier, T., Espinosa, B. and Scrivener, K.L. (1999) Quantitative Reitveldanalysis of calcium aluminate cement. World Cement 30(6).

French, P.J., Montgomery, R.G.J. and Robson, T.D. (1971) High strength concrete within the hour.Concrete August, 3.

George, C.M. (1976) The structural use of high alumina cement concrete. Revue des Materiaux etConstruction 201–209.

George, C.M. and Montgomery, R.G.J. (1992) Calcium aluminate cement concrete: durability andconversion. A fresh look at an old subject. Materiales de Construccíon 42, (228), 33.

Goyns, A. (2001) Calcium aluminate linings for cost-effective sewers. In Mangabhai, R.J. andGlasser, F.P. (eds), Calcium Aluminate Cements 2001. IOM Communications, London, pp. 617–631.

Lamour, V.H.R., Monteiro, P.J.M., Scrivener, K.L. and Fryda, H. (2001) Mechanical properties ofcalcium aluminate cement concretes. In Mangabhai, R.J. and Glasser, F.P. (eds), Calcium AluminateCements 2001. IOM Communications, London, 199–213.

Letourneux, R. and Scrivener, K.L. (1999) The resistance of calcium aluminate cements to acidcorrosion in wastwater applications. In Dhir, R.K. and Dyer, T.D. (eds), Modern ConcreteMaterials: Binders additions and admixtures. Thomas Telford, London, 275–283.

Mácias, A., Kindness, A., Glasser, F.P. (1996) Corrosion behaviour of steel in high alumina cementmortar cured at 5, 25 and 55°C: chemical and physical factors. J. Mat. Sci. 31, 2279–2289.

Majumdar, A.J. and Singh, B. (1992) Properties of some blended high-alumina cements. Cementand Concrete Research 22, 1101–1114.

Mangabhai, R.J. and Glasser, F.P. (eds) (2001) Calcium Aluminate Cements 2001. IOMCommunications, London.

Osborne, G.J. (1994) A rapid hardening cement based on high-alumina cement. Proc. Inst. Civ.Eng., Structures and Buildings 104, 93–100.

Calcium aluminate cements 2/31

Saucier, F., Scrivener, K.L., Hélard, L. and Gaudry, L. (2001) Calcium aluminates cement basedconcretes for hydraulic structures: resistance to erosion, abrasion & cavitation. Proc. 3rd Int.Conf. Concrete in Severe Conditions: Environment and Loading, CONSEC’01, Vancouver,Canada, June, Vol II, 1562–1569.

Scrivener, K.L., Lewis, M. and Houghton, J. (1997a). Microstructural investigation of calciumaluminate cement concrete from structures. Proc 10th ICCC, Göteborg, paper 4iv027.

Scrivener, K.L., Rettel, A., Beal, L. and Bier, T. (1997b) Effect of CO2 and humidity on themechanical properties of a formulated product containing calcium aluminate cement. Proc.IBAUSIL 1997, I-0745–0752, University of Weimar.

Scrivener, K.L. and Capmas, A. (1998) Calcium aluminate cements. In Hewlett, P.J. (ed.), F.M.Leas’s Chemistry of Cement and Concrete, 4th edn. Arnold, London, 709–778.

Scrivener, K.L., Cabiron, J-L. and Letourneaux, R. (1999) High performance concrete based oncalcium aluminate cements. Cem. Concr. Res. 29, 1215–1223.

Taylor, H.F.W., Famy, C. and Scrivener, K.L. (2001). Delayed ettringite formation, Cement andConcrete Research 31(5), 683–693.

Touzo, B., Glotter, A. and Scrivener, K.L. (2001) Mineralogical composition of fondu revisited. InMangabhai, R.J. and Glasser, F.P. (eds) Calcium Aluminate Cements 2001. IOM Communications,London, 129–138.

Zhou, Q. and Glasser, F.P. (2001) Thermal stability and decompositions mechanisms of ettringiteat <120°C. Cement and Concrete Research 31, 1333–1339.

PART 2

Cementitious additions

This Page Intentionally Left Blank

3

Cementitious additionsRobert Lewis, Lindon Sear,

Peter Wainwright and Ray Ryle

This chapter covers the production and use in concrete of the principal cementitiousadditions, namely, pulverized fuel ash and natural pozzolanas, ground granulated blastfurnaceslag, silica fume, metakaolin and limestone. It is recommended that this chapter be readin conjunction with Chapter 1.

3.1 The pozzolanic reaction and concrete

A pozzolana is a natural or artificial material containing silica in a reactive form. Bythemselves, pozzolanas have little or no cementitious value. However, in a finely dividedform and in the presence of moisture they will chemically react with alkalis to formcementing compounds. Pozzolanas must be finely divided in order to expose a largesurface area to the alkali solutions for the reaction to proceed. Examples of pozzolanicmaterials are volcanic ash, pumice, opaline shales, burnt clay and fly ash. The silica in apozzolana has to be amorphous, or glassy, to be reactive. Fly ash from a coal-fired powerstation is a pozzolana that results in low-permeability concrete, which is more durableand able to resist the ingress of deleterious chemicals.

3.2 Fly ash as a cementitious addition to concrete

The first reference to the idea of utilizing coal fly ash in concrete was by McMillan andPowers in 1934 and in subsequent research (Davis et al., 1935, 1937). In the late 1940s,

3/4 Cementitious additions

UK research was carried out (Fulton and Marshall, 1956) which led to the constructionof the Lednock, Clatworthy and Lubreoch Dams during the 1950s with fly ash as a partialcementitious material. These structures are still in excellent condition, after some50 years.

When coal burns in a power station furnace between 1250°C and 1600°C, theincombustible materials coalesce to form spherical glassy droplets of silica (SiO2), alumina(Al2O3), iron oxide (Fe2O3) and other minor constituents. When fly ash is added toconcrete the pozzolanic reaction occurs between the silica glass (SiO2) and the calciumhydroxide (Ca(OH)2) or lime, this is a by-product of the hydration of Portland cement.The hydration products produced fill the interstitial pores reducing the permeability ofthe matrix. Roy (1987) states ‘the reaction products are highly complex involving phasesolubility, synergetic accelerating and retarding effects of multiphase, multi-particle materialsand the surface effects at the solid liquid interface’. The reaction products formed differfrom the products found in Portland cement-only concretes. A very much finer porestructure is produced with time presuming there is access to water to maintain the hydrationprocess. Dhir et al. (1986) have also demonstrated that the addition of fly ash improvesthe dispersion of the Portland cement particles, improving their reactivity.

Figures 3.1 to 3.3 show that Ca(OH)2 (hydrated lime) is produced by the reaction ofthe Portland cement and water. Due to its limited solubility, particles of hydrated limeform within interstitial spaces. With a continuing supply of moisture, the lime reacts withthe fly ash pozzolanically, producing additional hydration products of a fine pore structure.The pozzolanic reaction is as in

Calcium hydroxide + silica = tricalcium silicate + water

3Ca(OH)2 + SiO2 = 3CaO · SiO2 + 3H2O

Figure 3.4, from Cabrera and Plowman (1981), shows the depletion of calcium hydroxide(lime) with time and how this reaction affects the long-term gain in strength of fly ashconcrete (A) compared to a PC concrete control (B). The reaction takes place both withinthe pores of the cement paste and on the surface of fly ash particles. Despite the pozolanic

PFAParticle

PFAParticle

PortlandCementGrain

PortlandCementGrain

Figure 3.1 Hydration products of Portland cement.

Cementitious additions 3/5

Figure 3.2 Lime is formed as a by-product of hydration.

PortlandCementGrain

PortlandCementGrain

PFAParticle

PFAParticle

Lime

Lime

Hydration products

Figure 3.3 The pozzolanic reaction products fill the interstitial spaces.

PFA reacts withlime filling poresin the matrix

PortlandCement

Grain

PortlandCementGrain

PFAParticle

PFAParticle

Lime

Lime

Strength

B

A Calciumhydroxidecontent

A

B

28 Age (days)28 Age (days)

Figure 3.4 Influence of calcium hydroxide on strength development (Cabrera and Flowman, 1981).

3/6 Cementitious additions

reaction reducing the available hydrated lime in the pore solution, there is sufficientremaining to maintain a high pH.

Little pozzolanic reaction occurs during the first 24 hours at 20°C. Thus for a givencementitious content, with increasing fly ash content, lower early strengths are achieved.Taylor (1997) explains the hydration processes involved in some detail. The presence offly ash retards the reaction of alite within the Portland cement at early ages. However,alite production is accelerated in the middle stages due to the provision of nucleation siteson the surface of the fly ash particles. The calcium hydroxide etches the surface of theglassy particles reacting with the SiO2 or the Al2O3–SiO2 framework. The hydrationproducts formed reflect the composition of the fly ash with a low Ca/Si ratio. Clearly thesurface exposed for reaction is greater, the finer the fly ash; additionally, the higher thetemperature, the greater the reaction rate. At later ages the contribution of fly ash tostrength gain increases greatly, provided there is adequate moisture to continue the reactionprocess.

3.2.1 The particle shape and density

Fly ash particles less than 50 μm are generally spherical (Figure 3.5) and the larger sizestend to be more irregular. The spherical particles confer significant benefits to the fluidityof the concrete in a plastic state by optimizing the packing of particles. The fly ashspheres appear to act as ‘ball bearings’ within the concrete reducing the amount of waterrequired for a given workability. In general the finer a fly ash the greater the water-reducing effect as in Figure 3.6 (Owens, 1979). Visually fly ash concrete may appear tobe very cohesive, until some form of compactive effort is applied. Any reduction in watercontent reduces bleeding and drying shrinkage.

Figure 3.5 Fly ash particles are spherical.

Cementitious additions 3/7

When relatively coarse fly ash, i.e. 45 μm residue >12 per cent, is interground withclinker or ground separately, the water requirement of concrete is markedly reducedaccording to Monk (1983). The grinding action appears to break down agglomerates andporous particles, but has little influence on fine glassy spherical particles, smaller thanapproximately 20 μm (Paya et al., 1996).

The particle density of fly ash is typically 2300 kg/m3, significantly lower than forPortland cement at 3120 kg/m3. Therefore, for a given mass of Portland cement a directmass substitution of fly ash gives a greater volume of cementitious material. The mixdesign for the concrete should be adjusted in comparison to a Portland cement of thesame binder content to allow for the increased volume of fine material.

For a given 28-day strength the higher cementitious content needed in comparisonwith Portland cement concrete and lower water content required for fly ash-based concretescan give significantly higher quality surface finishes.

3.2.2 Variability of fly ash

Controlled fineness fly ash has been available in the UK since 1975 and over recentyears, the majority of fly ash/PFA supplied for concrete complies with BS 3829 Part 1.This British standard limits the percentage retained on the 45 μm sieve to 12 per cent. Thetypical range of fineness found is < ± 5 per cent. This standard by default restricts thevariability of the PFA. In practice in order to achieve the required fineness, PFA to BS3892 Part 1 is classified using air-swept cyclones. The typical range of fineness found is± 3 per cent. BS EN 450 permits a wider range of variation of ± 10 per cent on thedeclared mean fineness range, which is up to a maximum of 40 per cent retained on the45 μm sieve.

The variability of the fineness may adversely influence the variability of the concrete,reducing the commercial viability of using fly ash. However, Dhir et al. (1981) concludedthat LOI and fineness are only ‘useful indicators’ of fly ash performance. Dhir shows thatthe differing characteristics of seven cements produce a greater influence on concrete

0 10 20 30 40 50 60

Fly ash residue at 45 microns (%)

50% PFA

40% PFA

30% PFA

20% PFA

10% PFA

Control 0% PFA in cement

220

210

200

190

180

170

160

150

Fre

e w

ater

con

tent

(lit

res/

m3 )

Figure 3.6 Finer fly ash and/or more fly ash reduces water content (Owens, 1979).

3/8 Cementitious additions

performance than the characteristics of eight differing fly ashes. He suggests that thefineness limits set in BS 3892 Part 1 are too low and that a limit of 20–25 per cent is morerealistic. Matthews and Gutt (1978) report on the effects of fineness, water reduction, etc.on a range of ashes with up to 18 per cent retained on the 45 μm sieve. At a fixed water/cement ratio, the range of strengths found indicates coefficients of variation of 17 percent at 365 days and 9 per cent at 28 days.

3.2.3 Fineness, water demand and pozzolanic activity

In one sense, fineness is important simply because a smaller particle size means a greatersurface area will be exposed to the alkaline environment within the concrete. With asingle source of fly ash, the strength performance can be related to water demand andfineness as in Figure 3.7, from Owens (1979).

Figure 3.7 Hydroxide concentration reduction due to pozzolanicity of various fly ashes.

Ash A unclassifiedAsh A classifiedAsh B unclassifiedAsh B classifiedAsh C unclassifiedAsh C classifiedPortland cementCalcium oxide saturation curve

35 40 45 50 55 60 65 70 75 80 85 90 95 100Hydroxide concentration (mmol/l)

18

16

14

12

10

8

6

4

2

0

Cal

cium

oxi

de c

once

ntra

tion

(mm

ol /

l) The calcium oxide concentrations produced by three ashes,both classified and unclassified when tested with onecement, relative to saturation level at that pH

BS EN 196-5 details a method of determining the pozzolanicity of a fly ash bycomparing the hydroxide concentration of a mortar containing fly ash against the theoreticalhydroxide content saturation possible. The results of a testing programme carried out1999 (NUSTONE Environmental Trust, 2000) clearly show the differences between sources(Figure 3.8) as well as the improved reactivity of classified fly ash.

The particle shape and finer fractions of fly ash are capable of reducing the watercontent needed for a given workability as in Figure 3.7. These effects are felt to be dueto void filling on a microscopic scale replacing water within the concrete mix. Dewar’s(1986) mix design system correlates with the water reductions found in practice whenusing fly ash. Where a water reduction is found this partially contributes to the relativestrength performance of the cement/fly ash combination by acting as a solid particulateplasticizer.

The strength of fly ash concrete will depend on whether a water reduction is achieved,plus the pozzolanic performance of the cement/fly ash combination. While finer fly ashes

Cementitious additions 3/9

are assumed to improve strength significantly with time, not all researchers find this true(Brown, 1980). Typically, strength development graphs of standard cubes with and withoutfly ash are shown in Figure 3.9. The graph also clearly illustrates what effect inclusion offly ash in concrete has on the development of strength at early ages. Where higher curingtemperatures are encountered, as in thick sections, significantly higher in-situ strengthcan be achieved than in test cubes cured at 20°C. Figure 3.10 shows 30 MPa grade,1.5 m cubic concrete specimens (Concrete Society, 1998), that have achieved considerablyhigher in-situ strengths than the 28-day, 20°C cube strengths. These were insulated onfive sides to recreate thick concrete sections.

70% CEM (PC) + 30% PFA

100% CEMI (Portland cement)

Test cubes cured at 20°C in water

Mixes designed to give equal28-day cube strengths

0 7 14 21 28 35 42 49 56 63 70 77 84 91Age (days)

1.4

1.2

1.0

0.8

0.6

0.4

0.2

0.0

Rel

ativ

e st

reng

th –

for

equ

al 2

8-da

y st

reng

th

Figure 3.8 At 20°C concrete containing 30 per cent fly ash continues to gain significant strength.

50% PC + 50% GGBS – 30 MPa concrete

100% Portland limestone cement –30-MPa concrete

100% Portland cement – 30 MPa

70% PC + 30% PFA – 30-MPa concrete

0 50 100 150 200 250 300 350 400

Age of concrete (days)

1.6

1.5

1.4

1.3

1.2

1.1

1

0.9

0.8

0.7

0.6Rel

ativ

e st

reng

th f

rom

cor

es/2

8-da

y cu

be s

tren

gth

Figure 3.9 30 per cent PFA can significantly improve the in-situ strength of concrete in the longer term.

3/10 Cementitious additions

3.2.4 Heat of hydration

Development of concrete mix design has seen an increase in the proportion of cementbeing replaced by fly ash. Fly ash is able to reduce the heat of hydration very effectively,as in Figure 3.10 by Woolley and Conlin (1989).

OPC/PFA – 265/150 kg/m3

OPC – 360 kg/m3

Relevant ambientprofiles

4 8 12 16 20 24 28 32 36 40

Relevant time from casting (hours)

60

50

40

30

20

10

Tem

pera

ture

(°C

)

Figure 3.10 Heat of hydration with time (Woolley and Coulin, 1989).

The hydration of Portland cement compounds is exothermic, typically with 500 Joulesper gram being liberated. The introduction of fly ash to replace a proportion of cement inconcrete influences temperature rise during the hydration period. However, the rate ofpozzolanic reaction increases with increasing temperature, however the peak temperaturesin fly ash concrete are significantly lower than equivalent PC concretes. Figure 3.11 byBamforth (1984) gives adiabatic temperature curves (°C per 100 kg of cement) for variouscement contents.

3.2.5 Setting time and formwork striking times

Using fly ash in concrete will increase the setting time compared with an equivalentgrade of PC concrete. There is a period before the hydration of fly ash concrete commences,but is has been shown by Woolley and Cabrera (1991) that the gain in strength, oncehydration has started, is greater for fly ash concrete. When 30 per cent fly ash is used toreplace PC in a mix, the setting time may be increased by up to 2 hours. This increasedsetting time reduces the rate of workability loss. However, it may result in finishingdifficulties in periods of low temperature. In compensation, it will reduce the incidenceof cold joints in the plastic concrete.

Formwork striking times at lower ambient temperatures may have to be extended incomparison to PC concrete, especially with thin sections. However, in practice verticalformwork striking times can be extended without this affecting site routines, e.g. theformwork is struck the following day. For soffit formwork, greater care has to be taken.

Cementitious additions 3/11

Reference should be made to BS 8110 (BSI) for recommended striking times. Temperature-matched curing can be used to ensure that sufficient in-situ strength has been achievedwhile allowing for the concrete curing conditions.

3.2.6 Elastic modulus

The elastic modulus of fly ash concrete is generally equal to or slightly better than thatfor an equivalent grade of concrete. Figure 3.12 shows the elastic modulus and strengthfrom Dhir et al. (1986) for concrete cured at different regimes.

15%

30%

50%

OPC

Cement content (kg/m3)300400500

15

14

12

10

8

6

4

2

0

–2

–4

Tem

pera

ture

ris

e (°

C p

er 1

00 k

g fly

ash

)

1 2 3 5 10 20 30 50 100 200 300 500Time (t + 1) hours

Figure 3.11 Temperature rise in fly ash and PC concretes (Bamforth, 1984).

Figure 3.12 Relationship between the modulus of elasticity and strength for fly ash and PC concretes (Dhiret al., 1986).

95% confidence limitsElastic modulas (KN/mm2)40

30

20

10

00 10 20 30 40 50 60 70 80 90

Compressive strength (N/mm2)

OPC PFA20°C water5°C water40°C airAir at 20°C/55% RH

3/12 Cementitious additions

3.2.7 Creep

The greater long-term strength for fly ash concretes gives lower creep values, particularlyunder conditions of no moisture loss. These conditions may be found in concrete remotefrom the cover zone of a structure. Where significant drying is permitted the strength gainmay be negligible and creep of OPC and fly ash concretes would be similar. Figure 3.13 showsthe creep of fly ash and OPC concrete loaded to different stress levels ((Dhir, et al., 1986).

3.2.8 Tensile strain capacity

Tensile strain capacity of fly ash concretes has been found to be marginally lower, andthese concretes exhibit slightly more brittle characteristics (Browne, 1984). There ispossibly a greater risk of early thermal cracking for given temperature drop, partiallyoffsetting the benefits of lower heat of hydration in the fly ash concrete.

3.2.9 Coefficient of thermal expansion

The type of coarse aggregate used largely influences the coefficient of thermal expansionof concrete. Replacement of a proportion of cement with fly ash will have little effect onthis property (Gifford and Ward, 1982).

3.2.10 Curing

Hydration reactions between cement and water provide the mechanism for the hardeningof concrete. The degree of hydration dictates strength development and all aspects of

OPCPFA

Creep (microstrain)Stress levels (%)

80

80

60

60

3030

0 60 120 180Time under load (days)

5000

4000

3000

2000

1000

0

Figure 3.13 Creep of fly ash and PC concrete loaded to different stress levels (Dhir et al., 1986).

Cementitious additions 3/13

durability. If concrete is allowed to dry out hydration will cease prematurely. Fly ashconcrete has slower hydration rates and the lack of adequate curing will affect the finalproduct. Thicker sections are less vulnerable than thin concrete sections because heat ofhydration will promote the pozzolanic reaction.

3.2.11 Concrete durability and fly ash

Keck and Riggs (1997) quote Ed Abdun-Nur who said, ‘concrete, which does not containfly ash, belongs in a museum’. They list the benefits of using fly ash, e.g. low permeability,resistance to sulfate attack, alkali–silica reaction and heat of hydration giving referencesto their conclusions. Thomas (1990) reported that cores taken from 30-year-old structuresin general have performed well; such structures must have been made with unclassifiedfly ashes over which minimal control of fineness, LOI, etc. would have been exercised.Berry and Malhotra review the performance of fly ash concretes in detail. They show thebeneficial properties of using fly ash for sulfate resistance, chloride penetration, permeability,etc. Consideration of some specific durability aspects follows.

Penetration of concrete by fluids or gases may adversely affect its durability. Thedegree of penetration depends on the permeability of the concrete, and since permeabilityis a flow property it relates to the ease with which a fluid or gas passes through it underthe action of a pressure differential. Porosity is a volume property, representing thecontent of pores irrespective of whether they are inter-connected and may/may not allowthe passage of a fluid or gas.

3.2.12 Alkali–silica reaction

Alkali–silica reaction (ASR) (Concrete Society Technical Report 30, 1999) is potentiallya very disruptive reaction within concrete. ASR involves the higher pH alkalis such assodium and potassium hydroxides reacting with silica, usually within the aggregates,producing gel. This gel has a high capacity for absorbing water from the pore solutioncausing expansion and disruption of the concrete. Some greywacke aggregates have beenfound particularly susceptible to ASR. The main source of the alkalis is usually the Portlandcement or external sources. Fly ash does contain some sodium and potassium alkalis butthese are mainly held in the glassy structure and not readily available. Typically, only some16–20 per cent of the total sodium and potassium alkalis in fly ash are water-soluble.

Many researchers have shown that fly ash is capable of preventing ASR. Oddly, theglass in fly ash is itself in a highly reactive fine form of silica. It has been found that theratio of reactive alkalis to surface area of reactive silica is important in ASR. A pessimumratio exists where the greatest expansion will occur. However, by adding more silica thatis reactive plus the effect of the dilution of the alkalis mean that disruptive ASR cannotoccur. The recommendations (BSI 5328 amendment 10365, 1999) within the UK requirea minimum of 25 per cent BS 3892 Part 1 fly ash to prevent ASR. For coarser fly ashes,a minimum of 30 per cent fly ash may be required to ensure sufficient surface area toprevent ASR. Small quantities of fine fly ash with low-reactivity aggregates and sufficientalkalis may be more susceptible to ASR if the pessimum silica–alkali ratio is approached.Even when total alkalis within the concrete are as high as 5 kg/m3, fly ash has been found(Alasali and Malhotra, 1991) able to prevent ASR. The addition of fly ash reduces the pH

3/14 Cementitious additions

of the pore solution to below 13 at which point ASR cannot occur. The use of low-alkalicements has a similar effect. However, the detailed mechanisms by which fly ash preventsASR are complex and imperfectly understood.

The ACI Manual of Concrete Practice (1994) suggests few restrictions on theeffectiveness of fly ashes. It states that ‘The use of adequate amounts of some fly ashescan reduce the amount of aggregate reaction’. It is later suggested that ashes only have tocomply with ASTM C618, which permits a wide range of fly ashes. Fournier and Malhotra(1997) investigated the ability of a range of fly ashes to prevent ASR. The activity index(AI) was found to affect the ASR performance. However, there was no correlation betweenfineness and AI. Nant-y-Moch, Dinas and Cwm Rheidol Dams are excellent examples offly ash preventing ASR constructed about the same time using the same aggregates. TheNant-y-Moch and Cwm Rheidol dams used ‘run of station’ fly ash as part of the cement.These dams have performed well, in comparison to the Portland cement-only Dinas dam,which has some evidence of ASR cracking.

3.2.13 Carbonation of concrete

As fly ash pozzolanically reacts with lime, this potentially reduces the lime available tomaintain the pH within the pore solution. However, fly ash does reduce the permeabilityof the concrete dramatically when the concrete is properly designed and cured. Whendesigning concretes for equal 28-day strength the slow reaction rate of fly ash usuallymeans that the total cementitious material is often increased. This increase partiallycompensates for the reduction in available lime. Coupling this with the lower permeabilityleads to the result that the carbonation of fly ash concrete is not significantly differentfrom Portland cement-only concrete of the same grade (28 days) (Concrete Society,1991). The ACI Manual of Concrete Practice (1994) confirms this view: ‘Despite theconcerns that the pozzolanic action of fly ash reduces the pH of concrete researchers havefound that an alkaline environment very similar to that in concrete without fly ash remainsto preserve the passivity of the steel’.

Carbonation is a complex function of permeability and available lime. With properlydesigned, cured and compacted fly ash concrete, carbonation is not significantly differentfrom other types of concrete. With extended curing and the low heat of hydration propertiesof fly ash concrete, the resulting low permeability may more than compensate for thereduced lime contents.

3.2.14 Sea water and chloride attack on reinforcing

Chlorides can come from many sources. First, it is important to restrict the chloridecontent in the constituents of the concrete to a minimum and guidance is given withinvarious standards, e.g. BS 5328, BS 8110 and BS 8500. Chlorides from external sourcesare normally from seawater and de-icing salts used on roads. Tri-calcium aluminate(C3A), a compound found in Portland cement, is able to bind chloride ions formingcalcium chloro-aluminate. Similarly, tetra calcium alumino ferrite (C4AF) can also reducethe mobility of chloride ions forming calcium chloro ferrite. Fly ash also contains oxidesof alumina, which are able to bind chloride ions.

Cementitious additions 3/15

One highly effective way of reducing chloride ingress is to lower the permeability ofthe concrete. When fly ash is added to concrete, the permeability is significantly lower asin Figure 3.14 from Owens (1979). This reduces the transportation rate of chloride ions,and many other materials, into the concrete.

Figure 3.14 By increasing the strength and fly ash content of concrete, the permeability reducessignificantly (Owens, 1979).

Estimate of chloride diffusion coefficient

25 MPa concrete

35 MPa concrete

50 MPa concrete

60 MPa concrete70 MPa concrete

0 10 20 30 40 50 60

% PFA content

100

10

1

0.1Log

scal

e –

diffu

sion

coe

ffici

ent

(D)

cm2 /s

× 10

9

3.2.15 Freeze–thaw damage

Fly ash concrete of the same strength has a similar resistance to freeze–thaw attack asPortland cement concrete. Dhir et al. (1987) reported on the freeze–thaw properties ofconcrete containing fly ash. They used nine fly ashes of various sieve residues. All mixeswere designed to give equal 28-day strengths. Freeze–thaw was assessed using an 8-hourcycle of 4 hours at –20°C and 4 hours at +5°C. Ultrasonic pulse velocity (UPV) andchanges in length were used for the assessment of freeze–thaw performance. They found,as have other researchers, that adding fly ash reduces freeze–thaw resistance unless airentrainment is used. However, Dhir found that only 1.0 per cent of air entrained in theconcrete gave superior performance compared with plain concrete irrespective of thecement type.

It is clear that the less permeable and denser the mortar matrix is within the concrete,the less space is available to relieve the pressures associated with the expansion offreezing water. The additional hydration products created by the pozzolanic reactionresult in fly ash being less frost resistant unless pressure-releasing voids are artificiallycreated by the addition of air entrainment. In all low-permeability concretes, irrespectiveof the fineness or chemistry of the fly ash or other constituents, freeze–thaw resistance isgiven by air entrainment. The air bubble structure is able to relieve expansive stresses ofthe freezing water presuming proper curing regimes have been adhered to.

3/16 Cementitious additions

3.2.16 Sulfate attack

Fly ash concrete can increase the resistance to sulfate attack compared with a CEM Iconcrete of similar grade. Deterioration due to sulfate penetration results from the expansivepressures originated by the formation of secondary gypsum and ettringnite. The beneficialeffects of fly ash have been attributed to a reduction of pore size slowing penetration ofsulfate ions. Less calcium hydroxide is also availble for the formation of gypsum.

The smaller pore size of fly ash concrete reduces the volume of ettringnite that may beformed. One of the major constituents of cement that is prone to sulfate attack, tricalciumaluminate (C3A), is diluted since a proportion of it will have reacted with the sulfateswithin the fly ash at an early age. Building Research Establishment Special Digest 1(BRE, 1991) discusses the factors responsible for sulfate and acid attack on concretebelow ground and recommends the type of cement and quality of curing to provideresistance. Concrete made with combinations of Portland cement and BS 3892: Part 1PFA, where the fly ash content lies between 25 per cent and 40 per cent has good sulfate-resisting properties.

Seawater contains sulfates and attacks concrete through chemical action. Crystallizationof salts in pores of the concrete may also result in disruption. This is a particular problembetween tide marks subject to alternate wetting and drying. Presence of a large quantityof chlorides in seawater inhibits the expansion experienced where groundwater sulfateshave constituted the attack.

3.2.17 Thaumasite form of sulfate attack

There has much been said about thaumasite in the UK since the discovery of a number ofdamaged structures on a UK motorway. Bensted (1988) describes the chemistry of thethaumasite reaction. Thaumasite attack occurs at temperatures below 15°C and requiresthe presence of calcium carbonate. Limestone aggregates are a source of calcium carbonate,with oolitic limestone appearing to be the most reactive form. Sulfates react with thecalcium carbonate and the C3S and C2S hydrates forming thaumasite. As these are thestrength-giving phases of the cement, their removal results in the concrete disintegratingto a white powdery sludge-like material. This reaction is not expansive and may not beeasily detected below the ground. At the time of writing, a number of research projectsremain outstanding, which should answer some of the questions this reaction poses.

Burton (1980) carried a series of mixes to determine the relative performance ofconcretes made with Portland cement, sulfate resisting Portland cement (SRPC) andPortland/fly ash blended cement concretes when exposed to sulfate attack. The sulfatesolutions were not heated and the tanks were in an unheated, external storeroom. Interestinglythe Portland cement mixes all showed signs of deterioration after six months in magnesiumsulfate solution but only slight deterioration in sodium sulfate solution. However, after 5years immersion total disintegration of all the PC samples occurred. The oolitic limestonemixes suffered between 22.2 and 42.5 per cent weight loss after one year in magnesiumsulfate and the crushed carboniferous limestone mixes 14.3 to 27.3 per cent weight loss.This indicates the effect of aggregate type on the rate of deterioration. Though thaumasitewas not well understood at the time of the project, a re-evaluation of the data and photographssuggests the deterioration found with both the PC and SRPC mixes was probably due tothe thaumasite form of sulfate attack.

Cementitious additions 3/17

Bensted believes the addition of fly ash should reduce the problems with thaumasitesimply because the permeability of fly ash concrete should prevent sulfate in solutionfrom penetrating to any depth.

3.2.18 Resistance to acids

All cements containing lime are susceptible to attack by acids. In acidic solutions, wherethe pH is less than 3.5, erosion of the cement matrix will occur. Moorland waters with lowhardness, containing dissolved carbon dioxide and with pH values in the range 4–7 maybe aggressive to concrete. The pure water of melting ice and condensation contain carbondioxide and will dissolve calcium hydroxide in cement causing erosion. In these situations,the quality of concrete assumes a greater importance.

3.3 Fly ash in special concretes

3.3.1 Roller-compacted concrete

High fly ash content roller-compacted concrete depends on achieving the optimum packingof all constituents in the concrete (Dunstan, 1981). That means that all voids should befilled. Optimizing the coarse aggregate content of a mix is common, but filling of thevoids in the mortar fraction is rarely considered. For mix design, the paste fraction is theabsolute volume of the cementitious materials and free water. The mortar fraction is theabsolute volume of the fine aggregate and the paste fraction. Adopting these mix designprinciples has resulted in roller-compacted concretes with 60–80 per cent of fly ash beingused in such applications.

3.3.2 High fly ash content concrete

Mix design principles developed for roller-compacted concrete have been used for concretecontaining high volumes of fly ash. These may be compacted by immersion vibrationplant (Cabrera and Atis, 1998). Adopting the principle of minimum voids in the paste,mortar and aggregate, mix designs have been successfully used for concrete placed tofloor slabs, structural basements and walls. These concretes have 40–60 per cant of thecementitious volume as fly ash. Other work, using the maximum packing, minimumporosity principle (Cabrera et al., 1984) have similarly been designed for structuralconcrete with fly ash making up to 70 per cent of the cementitious content by weight.

3.3.3 Sprayed concrete with pfa

Limited experimental work has shown that a major reduction in rebound can be achievedby replacing a proportion of cement with fly ash, or by using a blended fly ash–cementcomponent. Using the ‘Dry’ process of spraying and replacing 30 per cent of the cementcontent with fly ash, 30 per cent reduction of generated rebound was recorded. At the

3/18 Cementitious additions

same time, the hardened concrete revealed a smaller pore size with no loss of compressivestrength (Cabrera and Woolley, 1996). This finding has major potential for the replacementof a proportion of cement with pfa in sprayed concrete mixes.

3.4 Natural pozzolanas

Natural pozzolanas are glassy, amorphous materials that are normally classified into fourgroups:

1 Volcanic materials – these are rich in glass in an unaltered or partially altered state.They are found throughout the world in volcanic regions. They result from the explosionof magma that rapidly cooled by being quenched to form microporous glass. Oftenthese materials consist of >50 per cent silica, followed by alumina, iron oxides andlime. They often have a high alkali content of up to 10 per cent.

2 Tuffs – where a volcanic glass has been partially or fully transformed into zeoliticcompounds due to weathering. The range of the chemical composition is similar tovolcanic materials. The loss on ignition is an indication of the degree of transformationthat has occurred due to weathering.

3 Sedimentary – these are rich in opaline diatoms. The diatoms are the skeletons oforganisms that are mainly composed of opal. They have high silica content; however,these materials are often polluted with clay.

4 Diagenetic – These are rich in amorphous silica and derived from the weathering ofsiliceous rocks. They are normally high in silica and low in other oxides, depending onthe mineral composition of the parent rocks.

With all pozzolanic materials, the gradation and particle shape will have a significanteffect on the water requirement of the concrete. Naturally occurring pozzolanas oftenhave irregular shape that will increase water demand unless plasticizers are used tocompensate.

3.5 The use of ggbs in concrete

3.5.1 Introduction

The hydraulic potential of ground granulated blastfurnace slag (ggbs) was first discoveredas long ago as 1862 in Germany by Emil Langen. In 1865 commercial production oflime-activated ggbs began in Germany and around 1880 ggbs was first used in combinationwith Portland cement (PC). Since then it has been used extensively in many Europeancountries, such as Holland, France and Germany. In the UK the first British Standard forPortland Blastfurnace Cement (PBFC) was produced in 1923.

3.5.2 Production of ggbs

Blastfurnace slag is produced as a by-product during the manufacture of iron in a blastfurnace.It results from the fusion of a limestone flux with ash from coke and the siliceous and

Cementitious additions 3/19

aluminous residue remaining after the reduction and separation of the iron from the ore.To process the slag into a form suitable for use as a cementitious material one of twotechniques can be employed, namely granulation or pelletization; with either technique itis essential that the slag be rapidly cooled to form a glassy disordered structure. If the slagis allowed to cool too slowly this allows a crystalline well-ordered structure to formwhich is stable and non-reactive. With granulation the stream of molten slag is forcedover a weir into high-pressure water jets. This causes the slag to cool rapidly into glassygranules no larger than about 5 mm in diameter. In most modern granulators the watertemperature is kept below about 50°C and the water slag ratio is normally between tenand twenty to one. Following granulation the granulate is dried and ground to cementfineness in a conventional cement clinker grinding mill.

Pelletization involves pouring the molten slag onto a water-cooled steel-rotating drumof approximately 1 m diameter. The drum has fins projecting from it, which throw theslag through the air inside a building where water is sprayed onto it thus causing it to coolrapidly. Pelletizing produces material from about 100 mm down to dust; the larger particlestend to be crystalline in nature and have little or no cementitious value. Particles largerthan about 6 mm are therefore screened off and used as a lightweight aggregate inconcrete and only the finer fraction (< 6 mm) used for the manufacture of ggbs. Granulationis much more efficient at producing material with a high glass content, but the capitalcosts of a granulator are about six times greater than that of a pelletizer. Both materialscan be used as raw feed for ggbs and most Standards do not differentiate between them.

3.5.3 Chemical composition of ggbs

The chemical composition of slag will vary depending on the source of the raw materialsand the blastfurnace conditions. The major oxides exist within the slag glass (formed asthe result of rapid cooling) in the form of a network of calcium, silicon, aluminium andmagnesium ions in disordered combination with oxygen. The oxide composition of ggbsis shown in Table 3.1 where it is compared with that of Portland cement.

Table 3.1 Typical oxide composition of UKPortland cement and ggbs

Oxide Composition (%)Portland cement ggbs

CaO 64 40SiO2 21 36Al2O3 6.0 10Fe2O3 3.0 0.5MgO 1.5 8.0SO3 2.0 0.2K2O 0.8 0.7Na2O 0.5 0.4

3.5.4 Chemical reactions

In the presence of water ggbs will react very slowly but this reaction is so slow that onits own slag is of little practical use. Essentially the hydraulicity of the slag is locked

3/20 Cementitious additions

within its glassy structure (i.e. it possesses latent hydraulicity) and in order to release thisreactivity some form of ‘activation’ is required. The activators, which are commonlysulphates and/or alkalis, react chemically with the ggbs, and increase the pH of thesystem. Once a critical pH has been reached the glassy structure of the slag is ‘disturbed’,the reactivity is released and the slag will begin to react with the water producing its owncementitious gels. In practice activation is achieved by blending the ggbs with Portlandcement as the latter contains both alkalis (Ca(OH)2, NaOH and KOH) and sulphates. Thechemical reactions that occur between the Portland cement, the water and the ggbs arevery complex and are summarized below.

Hydration mechanism

• Primary reactions

OPC + water C.S.H.+ Ca(OH)

NaOH

KOH

2→ ⎫

⎬⎪

⎭⎪

(1)

GGBS + water + Ca(OH)2→C-(N,K)-S-H (2)NaOHKOH

• Secondary reactions:

(i) OPC primary reaction products + ggbs(ii) OPC primary reaction products + ggbs primary reaction products.

Reaction rateThe rate of the chemical reactions is influenced by several factors including:

• The quantity and relative proportions of the two components: the lower the proportionof cement, the slower the reaction.

• Temperature: just as with Portland cement the rate of the reaction increases with anincrease in temperature and vice versa

• The properties of the two components, in particular chemical composition and fineness.

Material propertiesThe most important properties of the slag are summarized in Tables 3.2 and 3.3 togetherwith the limiting values taken from the appropriate standard.

Chemical composition Various hydraulic indices have been proposed to relate compositionof the slag to hydraulicity; most of these imply an increase in hydraulicity with increasingCaO, MgO and a decrease with increasing SiO2. BS 6699 uses the term chemical modulusor reactivity index (see Table 3.2) and states that (CaO + MgO)/SiO2 should be greaterthan 1.0. In addition, as the CaO/SiO2 ratio increases, the rate of reactivity of the slag alsoincreases up to a limiting point when increasing CaO content makes granulation to a glassdifficult, BS 6699 limits this ratio to a maximum vlaue of 1.4

⎯⎯⎯⎯⎯→

Cementitious additions 3/21

Fineness Slag is normally ground finer than Portland cement to a fineness of> 400 m2/kg; the finer the material, the more reactive it becomes.

Glass countIt is the non-crystalline (glassy) component of the slag that is potentially reactive; thereforethe greater the proportion of glass, the more reactive the slag. The most common methodof measuring glass count is by X-ray diffraction (XRD) and when using this method theglass count should be ≥ 67 per cent.

A summary of some of the other physical properties of slag is shown in Table 3.3where comparisons are made with Portland cement. Because, like cement clinker, theslag granulate is ground its particle shape is angular although the surface texture tends tobe somewhat smoother than cement. The relative density of the slag is slightly lower thanthat of cement (2.9 cf 3.1).

3.5.5 Blended cement production

The blended cement produced from the combination of Portland cement with ggbs can bemanufactured by the intergrinding of the two components (i.e. clinker and granulatedslag) in the ball mills or by blending the two components (i.e. PC and ggbs) as separatepowers at the factory or by blending in the mixer. Intergrinding can lead to problemsbecause the slag granulate is much harder than the clinker leading to preferential grindingof the cement clinker. Most cement producers in Europe use separate grinding followedby factory blending of the two powders. In the UK at present the favoured method ofproduction is that of separate grinding combined with mixer blending.

The situation regarding Standards is a little confusing. There is a British Standard (BS6699) covering the powder (i.e. ggbs) but as yet no equivalent European Standard andthere is a European Standard (ENV 197-1 CEM II and CEM III) covering the slag cementwhich has replaced BS 146 and BS 4246.

Table 3.2 Limits for ggbs according to BS 6699

Property Limit

Chemical modulus/Reactivity index ≥ 1.0CaO + MgO

SiO 2

Lime–silica ratio CaOSiO 2

≤ 1.4

Glass count XRD ≥ 67%Fineness > 275 m2/kgMoisture content ≤ 1.0%

Table 3.3 Other properties of typical UK Portland cements andggbs

Property Portland cement ggbs

Shape Angular AngularRD 3.15 2.90Colour Grey White

3/22 Cementitious additions

Details on the limits of the two components (slag and clinker) for slag cements asgiven in ENV 197-1 are shown in Table 3.4. It can be seen from this table that it ispossible to replace up to 95 per cent of the cement with ggbs (CEMIII Type C) althoughreplacement levels of between 50–70 per cent are more typical for structural concreteapplications.

Figure 3.15 Influence of slag content and temperature on setting time (Concrete society, 1991).

25°C5°C

02550

% ggbs

0 5 10 15Time (hours)

4

3

2

1

0Pen

etra

tion

resi

stan

ce(N

/mm

2)

Table 3.4 European Standard ENV 197-1 for slag cements

Components CEM II CEM III(%) Portland slag cement Blastfurnace cement

Type A Type B Type A Type B Type C

PC clinker 80–94 65–79 35–64 20–34 5–19ggbs 6–20 21–35 36–65 66–80 81–95Minor constituents 0–5 0–5 0–5 0–5 0–5

3.5.6 Properties of concrete made with slag cements

Plastic concreteWater demand/workability For concrete made with equal slump, a lower water contentis required compared to Portland cement although the reductions are small and are nomore than about 3 per cent. This reduction is related largely to the smoother surfacetexture of the slag particles and to the delay in the chemical reaction.

Stiffening times Because the ggbs is slower to react with water than Portland cement itsuse is likely to lead to an increase in the stiffening times of the concrete. The extensionin stiffening time will be greater at high replacement levels (above about 50 per cent) andat lower temperatures (below about 10°C). Figure 3.15 (Concrete Society, 1991) showsthe influence of increasing additions of ggbs on the extension to setting times measuredin accordance with BS 4550.

Bleeding and settlement The influence of slag on bleeding and settlement depends tosome extent upon what basis comparisons are made. When testing at constant water/cement ratio the use of slag, in proportions greater than about 40 per cent, will invariably

Cementitious additions 3/23

Heat of hydration and early age thermal cracking The rate of heat evolution associatedwith ggbs is reduced as the proportion of slag is increased. This may be of benefit in largepours enabling greater heat dissipation and reduced temperature rise which will reducethe likelihood of thermal cracking. The actual temperature reductions that can be achievedin practice depend on many factors including: section size, cement content, proportion ofslag, and chemical composition and fineness of the cementitious components. The benefitsto thermal cracking from the reduced temperature rise may be offset to some extent by thelower creep characteristics of concretes containing ggbs (see below).

Figure 3.17 (Wainwright and Tolloczko, 1986) shows typical temperature–time profilesfor concretes with and without slag and Figure 3.18 (Bamforth, 1984) shows the influenceon temperature rise of different slag additions at different lift heights.

Hardened concreteCompressive strength and strength development Since ggbs hydrates more slowly thanPortland cement the early rate of strength development of slag concretes is slower, thehigher the slag content the slower the strength development. However, provided adequatemoisture is available, the long-term strength of the slag concretes is likely to be higher.This higher later age strength is due in part to the prolonged hydration reaction of the slagcements and to the more dense hydrate structure that is formed as a result of the slowerhydration reaction.

As with Portland cements the rate of the hydration reaction of slag cements is temperaturedependent but slag cements have a higher activation energy than Portland cements andtherefore their reaction rates are more sensitive to temperature change. This means that asthe temperature increases the increase in the rate of gain of strength of slag cement isgreater than that of Portland cement but the opposite is true as the temperature reduces.The influence of temperature on strength development is of significance when consideringthe behaviour of concrete in situ. In such situations the rate of strength development and

ggbs%0

4070

10 20 30 40 50 6028-day compressive strength (MPa)

50

40

30

20

10

0Ble

edin

g ra

te (

ml/c

m2/s

)×

106

Figure 3.16 Influence of ggbs on bleeding at equal compressive strengths (Wainwright, 1995).

lead to increases in both bleeding and settlement. When, however, comparisons are madeon the basis of equal 28-day strength the differences are not so marked as can be seen inFigure 3.16 (Wainwright, 1995). It should also be noted that there are marked differencesbetween the bleed characteristics of Portland cements from different sources and in somecases these differences can be as great as those resulting from the addition of slag.

3/24 Cementitious additions

ultimate strength may be appreciably different from that indicted by standard curedcubes. In thick sections or cement-rich concrete the early temperature rise may well be inexcess of 50°C as shown previously in Figure 3.3. The effect of this is to accelerate theearly strength gain and to impair long-term development for concretes with and withoutggbs. However, the amount by which the long-term strength gain is reduced is lower forslag cements than for Portland cements. Conversely in thin sections cast in cold weather

10 100

Log time (hours)

Start Temp Time toOPC Slag temp °C rise °C peak hours

100 0 20.5 45.5 7150 50 22.2 41.6 13830 70 20.9 35.6 170

50

40

30

20

10

0

Tem

pera

ture

ris

e ( °

C)

%

Figure 3.17 Influence of slag content on adiabatic temperature rise (Wainwright and Tolloczko, 1986).

Lift height (m)4.5

3

2

1.5

1

10 20 30 40 50 60 70 80

Replacement level (%)

100

90

80

70

60

50

% o

f te

mpe

ratu

re r

ise

in P

C c

oncr

ete

Figure 3.18 Influence of ggbs on temperature rise for various lift heights (Bamforth, 1984).

Cementitious additions 3/25

conditions the slower rate of gain of strength may lead to some extension in formworkstriking times.

The influence of temperature on strength development is shown in Figure 3.19(Wainwright and Tolloczko, 1986) and Figure 3.20 (Swamy, 1986). Concretes whichhave been subjected to ‘heat cycled’ curing in Figure 3.19 have undergone a temperaturecycle similar to that which they are likely to experience in a structure with relatively thicksections (i.e. similar, for example, to that shown in Figure 3.17). It is clear from Figure3.19 that under heat cycled conditions the 70 per cent slag mix exceeded the strengthof the PC mix after about 3 days yet under normal curing conditions it was still about10 N/mm2 below the PC mix after 6 months. In Figure 3.20 the specimens have been keptunder constant temperature conditions, the mixes are not the same as those shown inFigure 3.19. Figure 3.20 illustrates how much more the strength of the slag mix isretarded at 5°C compared to the PC mix

Normal curingHeat cycled curing

80

70

60

50

40

30

20

10

Com

pres

sive

str

engt

h (N

/mm

2 )

1 2 3 7 28 56 168

1 2 3 7 28 56 168Age (days) (log scale)

OPC Slag

100 050 5030 70

%

Figure 3.19 Comparison of strength development of concretes with and without slag, normal cured andheat cycled cured (Wainwright and Tolloczko, 1986).

Tensile strength Compared to Portland cement slag cement concretes tend to have aslightly higher tensile strength for a given compressive strength although the differencesare of little practical significance.

Elastic modulus The use of ggbs will have the effect of increasing slightly the elasticmodulus of the concrete for a given compressive strength when compared with Portlandcement. However, as with tensile strength the magnitude of this difference is not large

3/26 Cementitious additions

50/50OPC/slag

Temp. °C30

20

5

0 1 2 3 7 16 28

1 2 3 7 16 28

Age (days) (log scale)

Cement content 360 kg/m3

Water: cement ratio 0.40

OPC

60

50

40

30

20

10

Com

pres

sive

str

engt

h (N

/mm

2 )

Figure 3.20 Influence of temperature on strength development for concretes with and without slag(Swamy, 1986).

and the relationship between strength and modulus for all both slag and Portland cementsfollows closely that given in BS 8110.

Creep Under conditions of no moisture loss the basic creep of concrete is reduced as thelevel of ggbs is increased. For high replacement levels (i.e. > 70 per cent ggbs) reductionsas high as 50 per cent have been reported. The reductions in creep are generally associatedwith the greater strength gain at later ages of concretes containing slag. Under dryingconditions this later age strength gain may be small and the differences in creep are likelyto be less pronounced. For most practical situations where drying shrinkage is moderatethe behaviour of concretes made from slag cements is likely to be similar to that ofPortland cement concretes. Figures 3.21 and 3.22 (Neville and Brooks, 1975) show theinfluence of ggbs additions on both the basic and drying creep of concrete.

Figure 3.21 Effect of ggbs on basic creep in water at 22°C (Neville and Brooks, 1975).

Replacement:0

30%

50%

0 50 100 150 200

Time under load (days)

400

200

0

Bas

ic c

reep

, 10

–6

Cementitious additions 3/27

Drying shrinkage The use of ggbs has very little if any influence on the drying shrinkageof the concrete.

Other properties related to constructionSurface finish The slight improvement in workability and small increase in paste volumeof ggbs concretes generally makes it easier to achieve a good surface finish. In additionthe colour of the concrete will be lighter than that of Portland cement concretes.

Formwork pressures As a result of the extended setting times mentioned earlier the useof slag in concrete will lead to an increase in formwork pressures (Clear and Harrison,1985). The increase in pressure will be particularly pronounced in high lifts (> 4 m) castat low temperatures (< 5°C) and at low placing rates (< 0.5 m/h).

Formwork striking times The slower rate of gain of early strength of concretes containinghigh levels of ggbs may require the extension of formwork striking times. In practice,however, the actual construction process often requires concrete to be cast one day andvertical formwork struck the next. In such cases it is quite likely that the minimumstriking times will in any case be extended and that therefore the use of slag may notaffect the actual construction process.

Curing In situations where durability is recognized as a potential problem then it maybe expected that concretes containing slag will require longer curing periods particularlyin cold weather conditions and where thin sections are involved.

DurabilityIt is generally recognized that the durability of concrete is primarily related to its permeability/diffusion to liquids and gases and its resistance to penetration by ions such as sulfates andchlorides. Generally speaking, provided the concretes have been well-cured slag concretesare likely to be more durable than similar Portland cement concretes.

Replacement:30%

50%

0

0 50 100 150 200

Time under load (days)

2000

1000

Tota

l cre

ep,

10–6

Figure 3.22 Effect of ggbs on total creep at 21°C and 60 per cent RH (Neville and Brooks, 1975).

3/28 Cementitious additions

Permeability In well-cured concretes the inclusion of ggbs will be beneficial in terms ofthe long-term permeability and particularly so at higher temperatures. The reasons forthis are:

1 The continued hydration which takes place well beyond 28 days in the slag concretes.2 The overall finer pore structure associated with ggbs as shown in Figure 3.23 (Pigeon

and Regourd, 1983), as the slag content increases, the number of smaller pores increasesand the overall pore size distribution becomes finer even though there may be littlechange in porosity.

3 The influence of elevated temperature on pore structure is illustrated in Table 3.5(Kumar et al., 1987). In these tests conducted on cement pastes an increase in curingtemperature resulted in a coarsening of the pore structure for Portland cements yet forthe slag cements the pore structure was little affected.

Figure 3.23 Influence of slag content on pore size distribution of pastes (Pigeon and Regourd, 1983).

0% slag

30% slag

60% slag

10–3 10–2 10–1 100 101 102

Pore radius (μm)

25

20

15

10

5

0

15

10

5

0

10

5

0

Pro

port

ion

of t

otal

por

e vo

lum

e (%

)

Figure 3.24 (Grube, 1985) illustrates the problems associated with trying to relatesmall-scale laboratory tests data to full scale in-situ situations where the effects of temperatureand the reserve of moisture may be significant. The results are from cast cylinders andfrom cores taken from walls 200 mm thick. In both cases the concrete was made to thesame water/cement +ggbs ratio and storage conditions were either outdoors or in laboratoryair at 20°C, 65 per cent RH. After only three days curing the cores made with the slagconcrete exhibited lower permeabilities to oxygen than those made from OPC yet withthe cast specimens even after 28 days the slag had only just reached parity with the PC.

Cementitious additions 3/29

Carbonation The ability of the concrete to protect the steel from corrosion depends,among other factors, upon the extent to which the concrete in the cover zone has carbonated.The influence of additions of ggbs on carbonation has been the subject of much researchand there still appears to be some disagreement as to its effects. The reasons for much ofthis debate appear to be related to the test procedures and conditions used in the studiesand to the basis on which comparisons have been made.

Results obtained on cubes subjected to outdoor exposure conditions (Figure 3.25,Osborne 1986) indicate that, on the basis of equal 28-day compressive strength, there isa marginal increase in carbonation depths for ggbs concrete but for strengths above about

Table 3.5 Influence of temperature and cement type on pore structure of cement pastes

Porosity (%) Medium pore size (nm)Cured at Cured at

Composition Age 27°C 38°C 60°C 27°C 38°C 60°C(by wt) (days)

Type I Portland 7 25.0 24.0 25.0 13.5 12.0 9.0cement (OPC) 28 20.0 22.0 22.0 13.0 15.0 13.5

90 17.0 19.0 24.5 10.5 12.0 17.5

65% GGBS 7 21.5 19.0 17.0 13.5 13.5 11.0+ 28 12.2 11.5 12.0 2.75 2.75 2.9535% OPC 90 9.5 10.0 10.0 2.30 2.45 2.60

180 8.0 8.0 8.0 2.35 2.40 2.50

Cores from walls

1 3 10 281 3 10 28

Shutter stripping time (days)

OPC

GGBS

Permeability (m2)

10–15

10–16

10–17

10–18

Shutter stripping time (days)

OPC

GGBS

Permeability (m2)

10–15

10–16

10–17

10–18

Cast specimens

Figure 3.24 Relationship between permeability and stripping time for Portland cement and slag cementconcretes (Grube, 1985).

3/30 Cementitious additions

30.0 N/mm2 the differences are unlikely to be of practical significance. In other tests inwhich concretes were exposed at an early age (< 24 hours) the depth of carbonation wasfound to be largely controlled by the water/Portland cement ratio (Figure 3.26, ConcreteSociety, 1991). In such cases the depth of carbonation of the slag concretes was found tobe significantly higher particularly at water/cement ratios above about 0.45. In contrast,extensive surveys of old concrete structures carried out on the Continent have shown nodifferences in carbonation depths for concretes with and without slag. Similar surveyscarried out in the UK found an increase in carbonation depth at high replacement levels(> about 65 per cent) where the concrete was exposed to a sheltered environment butwhere the concrete was exposed to rainfall there was hardly any difference.

(80)

(60)

(50)

(40)

(0)

9-year slag9-year OPC (different

test series)

⎫⎬⎭

( ) = % slag

15 20 25 30 35 40 45 50 55

28-day water cured compressive strength (N/mm2)

40

30

20

10

0

Car

bona

tion

dept

h (m

m)

Figure 3.25 Influence of slag on relationship between carbonation depth and strength (Osborne, 1986).

Figure 3.26 Relationship between depth of carbonation of ggbs and OPC concretes and water/(OPC +ggbs) ratio (Concrete Society, 1991).

OPC

ggbs

Depth of carbonation (mm)

0 0.2 0.4 0.6 0.8 1.0 1.2

Water/(OPC + ggbs) ratio

70% ggbs

50% ggbs

20

15

10

5

0

Cementitious additions 3/31

Alkali–silica reaction (ASR) The alkali–silica reaction is the most common form ofalkali aggregate reaction and occurs as the result of a reaction between the alkaline porefluid in the cement and siliceous minerals found in some aggregates. This reaction resultsin the formation of a calcium silicate gel which imbibes water producing a volumeexpansion and disruption of the concrete.

One effective way of reducing this expansion is by the use of ggbs as illustrated Figure3.27 (Hobbs, 1982) which shows results of a laboratory study carried out in the UK.Although ggbs contains alkalis (in some cases quite a high level) their solubility issomewhat lower than those found in Portland cement. The mechanism by which thesealkalis contribute to the alkalinity of the pore solution and therefore take part in thereaction is complex and is still not yet fully understood. In 1999 the Building ResearchEstablishment in the UK published modified guidelines in the form of a Digest (BREDigest 330, 1999) on ways to minimize the risk of attack from alkali–silica reaction. TheDigest is a comprehensive document published in four parts and it is only intended hereto highlight the main points contained in it as outlined below:

• Account must be taken of the reactivity of the aggregate, the aggregates being classifiedas: Low, Normal and High Reactivity.

• The use of a low alkali cement is recommended which is defined as a cement with analkali content of ≤ 0.6 per cent Na2O equivalent.

Figure 3.27 Influence of slag content on expansion due to ASR (Hobbs, 1982).

0 100 200 300Time (days)

50

10

30

40

50

% slag02.0

1.5

1.0

0.5

0

Exp

ansi

on (

%)

3/32 Cementitious additions

• One way of achieving a low alkali cement is by blending Portland cement with ggbs,the minimum recommended levels of slag being, 50 per cent or 40 per cent ggbsdepending on aggregate reactivity.

• Where the minimum level of ggbs is used, no account need be taken of the alkalicontent of the slag when calculating the total alkali content of the concrete.

• The limit on total alkalis in the concrete ranges from 2.5–5.0 kg Na2O equiv/m3

depending on the reactivity of the aggregate.

The procedures for calculating the total alkali content in the concrete are far morecomplex than those given in previous guidelines. Readers are strongly advised to consultthe Digest before attempting to carry out any calculations.

Sulfate resistance Concretes containing ggbs are acknowledged to have higher resistanceto attack from sulfates than those made with only Portland cements. This improvedresistance is related to the overall reduction in the C3A content of the blended cement(ggbs contains no C3A) and to the inherent reduction in permeability. It is generallyaccepted that provided the Al2O3 content of the slag is less than 15 per cent then cementscontaining at least 70 per cent ggbs can be considered comparable to sulfate-resistingPortland cement.

In the UK recommendations relating to resistance to sulfate attack have been revisedfollowing the publication in 2001 of BRE Special Digest 1 ‘Concrete in AggressiveGround’ (BRE, 2001). This document replaces BRE Digest 363 ‘Sulfate and Acid Resistanceof Concrete in the Ground’ and was written partly as a result of the relatively recentdiscovery of the thaumasite form of sulfate attack found in the foundations of somestructures in the UK. The document is in four parts and is far more comprehensive andcomplex than its predecessor. It coveres all forms of aggressive ground conditions, includingsulfates, acids, thaumasite and brownfield sites. Readers will need to consult this documentfor details of the procedures that have to be followed and because of the complexity ofthese procedures and the number of variables involved it is difficult to give simpleexamples to illustrate the recommendations resulting from guidelines. If, for example, thesoil was classified as Design Class 3*(DC3) then, depending on the aggregate type andother conditions, the requirements might be as shown in Table 3.6.

Table 3.6

Cement type Min cement Max(kg/m3) w/c

OPC + � 70% � 90% ggbs 380 0.45orsulfate-resisting Portland cement

Chloride ingress Slag cements are shown significantly more resistant to the ingress ofchloride ions than Portland cements. The reason for this it is not simply due to thereduced permeability of the slag cements but also because the chlorides chemicallycombine with the slag hydrates which has the effect of reducing the mobility of the

*DC3 = 1.5 – 3.0 g/l SO4 in groundwater which is equivalent to 1.2 – 2.5 g/lSO3 as defined in Digest 363 and BS 8110

Cementitious additions 3/33

chlorides. This improved resistance has the potential for minimizing the risk of corrosionof the reinforcement in concrete.

Results in Figure 3.28 (Smolczyk, 1977) show that even at a high water/cement ratioof 0.7 the concrete containing slag had a significantly lower chloride diffusion rate thaneven the Portland cement concrete with the lowest water/cement ratio of 0.5.

Figure 3.28 Influence of slag content on chloride diffusion (Smolczyk, 1977).

100% OPC

40% slag

70% slag

1 year 2 years

100% OPC

w/c = 0.70.60.5

5

4

3

2

1

0

Chl

orid

e co

nten

t (%

cem

ent)

Results of a more recent long-term study using both laboratory specimens and specimenstaken from real structures are summarized in Figure 3.29 (Bamforth, 1993). These datashow that there is a significant reduction with time in the chloride diffusion coefficientsof concretes containing ggbs whereas with Portland cement concretes there is very littlechange.

OPC

ggbs

0.1 1 10 100

Time (years)

E-10

E-11

E-12

E-13

E-14

Diff

usio

n co

effic

ient

(m

2 /s)

Figure 3.29 Change of effective chloride diffusion coefficient with time for concretes with and withoutggbs (Bamforth, 1993).

3/34 Cementitious additions

Alkalinity Electrochemical measurements have shown that despite the reduction incalcium hydroxide, caused by the secondary reactions of the ggbs hydration, the pH ofthe paste remains at a level which is well in excess of that which would affect thepassivity of the reinforcing steel.

Freeze–thaw resistance There is little difference between the freeze–thaw resistance ofPortland cement and ggbs concretes of a similar strength and air content. However, non-air-entrained concretes with high slag levels (> 60 per cent) are likely to show an inferiorresistance to attack than equivalent Portland cement concretes. The use of ggbs also doesnot normally impair the effectiveness of air-entraining admixtures to entrain air.

Abrasion resistance Provided the concrete has been adequately cured and when comparisonsare made on the basis of equal grade there would appear to be a slight advantage in termsof abrasion resistance from the use of ggbs. However, under conditions of inadequatecuring the abrasion resistance of all concretes will be significantly reduced although ggbsconcretes appear to be more affected than those made from Portland cement.

3.5.7 Concluding remarks

Slag concretes have been used successfully in concrete for many yeras in many countriesthroughout the world. There is much evidence, both from laboratory test data and inservice records, to suggest that there are potentially many technical benefits to be gainedfrom using this material. Many countries have accepted the benefits and have recommendedits use in their national standards.

Provided the user is made aware of the properties of the material and has therefore anunderstanding of both the advantages and disadvantages to be gained, there is no reasonwhy it should not continue to be used successfully and more often in the future.

3.6 Silica fume for concrete

3.6.1 The material

The terms condensed silica fume, microsilica, silica fume and volatilized silica are oftenused to describe the by-products extracted from the exhaust gases of silicon, ferrosiliconand other metal alloy smelting furnaces. However, the terms microsilica and silica fumeare used to describe those condensed silica fumes that are of high quality, for use in thecement and concrete industry. The latter term, silica fume, will be the one used in the newEuropean Standard – prEN 13263-1.

Silica fume was first ‘obtained’ in Norway, in 1947, when environmental restraintsmade the filtering of the exhaust gases from the furnaces compulsory. The major portionof these fumes was a very fine powder composed of a high percentage of silicon dioxide.As the pozzolanic reactivity for silicon dioxide was well known, extensive research wasundertaken, principally at the Norwegian Institute of Technology. There are over 3000papers now available that detail work on silica fume and silica fume concrete.

Large-scale filtering of gases began in the 1970s and the first standard, NS 3050, foruse in a factory-produced cement, was granted in 1976.

Cementitious additions 3/35

Production and extractionSilica fume is produced during the high-temperature reduction of quartz in an electric arcfurnace where the main product is silicon or ferrosilicon. Due to the vast amounts ofelectricity needed, these furnances are located in countries with abundant electrical capacityincluding Scandinavia, Europe, the USA, Canada, South Africa and Australia.

High-purity quartz is heated to 2000°C with coal, coke or wood chips as fuel and anelectric arc introduced to separate out the metal. As the quartz is reduced it releasessilicon oxide vapour. This mixes with oxygen in the upper parts of the furnace where itoxidizes and condenses into microspheres of amorphous silicon dioxide.

The fumes are drawn out of the furnace through a precollector and a cyclone, whichremove the larger coarse particles of unburnt wood or carbon, and then blown into a seriesof special filter bags.

The chemistry of the process is very complex and temperature dependent. The SiCformed initially plays an important intermediate role, as does the unstable SiO gas whichforms the silica fume:

At temperatures > 1520°CSiO2 + 3C = SiC + 2CO↓↓At temperatures > 1800°C3SiO2 + 2SiC = Si + 4SiO + 2CO

The unstable gas travels up in the furnace where it reacts with oxygen to give the silicondioxide:

4SiO + 2O2 = 4SiO2

The silica fume is not collected as pure material as other particles and chemicals arepresent in the powder collected in the filter bags (See Figures 3.30 and 3.31)

CharacteristicsSilica fume is, when collected, an ultrafine powder having the following basic properties:

1 At least 85 per cent SiO2 content.2 Mean particle size between 0.1 and 0.2 micron.3 Minimum specific surface of 15 000 m2/kg.4 Spherical particle shape.

The powder is normally grey in colour but this can vary according to the source. Operationof the furance, the raw materials and quality of metal produced, will all have an effect onthe colour of the powder. This variation can show up in the material from one furnace aswell as from different furnaces, thus it is advisable to ensure a consistency of supply fromone source (Table 3.7).

Health and safetyAs with all fine powders there are potential health risks, particularly in relation to silicondioxide and the lung disease silicosis. In studies of the material and of workers in theferrosilicon industry (Aitcin and Regourd, 1985; Jorgen, 1980) it has been found that the

3/36 Cementitious additions

Figure 3.30 Typical plant layout: (left to right): furnace, main stack, pre-collector, cooling pipes and fan,and baghouse. (Courtesy Elkem Materials Ltd).

0.5 μmMicrosilica

Figure 3.31 Undensified silica fume (SEM).

Cementitious additions 3/37

silicon dioxide that causes silicosis is the crystalline form. Silica fume, produced asdescribed above, is amorphous and has been found to be non-hazardous.

The threshold limit values for the respirable dust have been set at 3 mg/m3 (low) and5 mg/m3 (high) and exposure to concentrations above the high limit are not advised. TheCAS number for both powder and slurry is:

Amorphous SiO2: 7631-86-9.

For full health and safety information it is always advised to contact the producers.

Available forms of silica fumeAs the powder is a hundred times finer than ordinary Portland cement there are transportation,storage and dispensing considerations to be taken into account. To accommodate some ofthese difficulties the material is commercially available in various forms. The differencesbetween these forms are related to the shape and size of the particles and do not greatlyaffect the chemical make-up or reaction of the material. These differences will influencethe areas of use and careful thought should be given to the type of silica fume chosen fora specific application. The main forms are as follows.

Undensified

Bulk density: 200–350 kg/m3

Due to the very low bulk density, this form is considered impractical to use in normalconcrete production. The main areas of use are in refractory products and formulatedbagged materials such as grouts, mortars, concrete repair systems and protective coatings.

Densified

Bulk density: 500–650 kg/m3

In the densification process the ultrafine particles become loosely agglomerated, makingthe particulate size larger. This makes the powder easier to handle, with less dust, than theundensified form. Areas where this material is used are in those processes that utilize highshear mixing facilities such as pre-cast works, concrete rooftile works or readymixedconcrete plants with ‘wet’ mixing units.

Table 3.7 A comparison of cementitious materials available in the UK

Portland cement pfa ggbfs Microsilica

Physical data for cementitious materialsSurface area (m2/kg) 350–500 300–600 300–500 15 000–20 000Bulk density (kg/m3 1300–1400 1000 1000–1200 200–300Specific gravity 3.12 2.30 2.90 2.20

Chemical data for cementitious materialsSiO2 20 50 38 92Fe2O3 3.5 10.4 0.3 1.2Al2O3 5.0 28 11 0.7CaO 65 3 40 0.2MgO 0.1 2 7.5 0.2Na2O + K2O 0.8 3.2 1.2 2.0

3/38 Cementitious additions

Micropelletized

Bulk density: 700–1000 kg/m3

Micropelletization involves forming the powder into small spheres about 0.5–1 mm indiameter. The material in this form does not readily break down in conventional concretemixing and is best suited to intergrinding with cement clinker to produce a compositecement. Icelandic cement is made in this fashion and contains 7.5 per cent silica fume tocombat the potential ASR from the local aggregates.

Slurry

Specific gravity: 1400 kg/m3

This material is produced by mixing the undensified powder and water in equal proportions,by weight, to produce a stable slurry.

Mixing and maintaining a stable slurry requires expensive hi-tech equipment andcannot be done easily and, therefore, all slurries shold be obtained from one specificsupplier to maintain quality. In this form the material is easily introduced into the concretemix. It is also the most practical form for dispensing by weight or volume. It can be usedfor virtually all forms of concrete from semi-dry mixes for pre-cast products to self-compacting concrete and is ideally suited to the readymixed industry.

3.6.2 Inclusion in concrete

Standards and specificationsVarious national standards, codes of practice and recommendations have been publishedsince the use of silica fume has become globally accepted (See Table 3.8). The current listis as follows:

• Arabian Gulf – Rasheeduzzafar et al. Proposal for a code of practice to ensure durabilityof concrete construction in the Arabian Gulf environment. King Faud University,Saudi Arabia.

• Australia – AS 3582.3 1994• Canada – CAN/CSA-A23.5-M98. Supplementary cementing materials.• China – draft• Denmark – DS 411. Code of practice for the structural use of concrete.

The Danish Academy of Technical Sciences Hefte no. 25.Basic Concrete Specification. Copenhagen, May 1986

• Finland – Suomen Betoniyhdistys r.y. Betoninormit 1987, RakMK BY, by 15, Helsinki• Germany – Institut Für Bautechnik. PA VII-21/303• India – draft (IS 1727)• Japan – JIS A6207:2000• New Zealand• Norway – NS 3098. Portland cements, specification of properties, sampling and delivery.

NS 3420. Specification texts for building and construction, NBR, Oslo• Sweden – Statens Planverk PFS 1985:2. Mineral additions to concrete, approval rules.

Staten Planverk, Stockholm

Cementitious additions 3/39

• UK – Agrément Certificate No. 85/1568. British Board of Agrément• USA – ASTM C-1240.xx. (always use newest version)• EU – prEN 13263-1 (expected vote 2003)

Effects on fresh concreteDue to the nature and size of the silica fume, a small addition to a concrete mix willproduce marked changes in both the physical and chemical properties. The primaryphysical effect is that of adding, at the typical dosage of 8–10 per cent by cement weight,

Table 3.8 Standards comparison table

USA Norway Canada Australia Europe France JapanASTM NS 3045 CSA, AS 3582 CEN prEN NF P JISC1240:01 A23.5-98 13263-1 18-502 A6207:

(Provisional) 1992 2000

SiO2 %> 85.0 85 85 85 85 85 85SO3 %< 1.0 3 2.0 2.5 3.0Cl %< Report Report 0.3 0.2 0.1

if >0.10CaO %< 2 1.0 1.2 1.0MgO 5.0Si (free) %< 0.4 0.4Total alkalis 4Free C 4Moisture 3.0 2 1 3.0content %<LOI %< 6.0 5 6.0 6 4.0 5.0Specific >12 15–35 20–35 >15surfacem2/g >Bulk density Report ReportPozzolanic 85% – 7d 95% – 85% – 7d Report 100% 28d 95% – 7dactivity accel’d 28d accel’d Normal 105% –index, %> curing, Normal curing curing 28d

w/cm = curing w/p ratio w/c = 0.5variable w/c = 0.5

= 0.5Retained on 10 1045 micronsieve % <Density,kg/m3 2100– Report

2300Autoclave 0.2%expansion % <Canadian No visibleFoaming test foamNotes Character- Characteristic Material

istic values. Not type Avalues an official

standard, inthe approvalprocess.

Note: The main standards that are accepted on a global basis are the ATSM C1240 and the CSA A23.5. Onlythe latest versions of any of these standards should be accepted for specification use. The table gives only themandatory chemical and physical requirements. Several of the standards also contain optional requirements.Where there are blanks, no mandatory requirements exist.

3/40 Cementitious additions

between 50 000 and 100 000 microspheres per cement particle. This means that the mixwill be suffused with fine material causing an increase in the cohesiveness of the concrete.When using a powder form of silica fume this will mean an increased water demand tomaintain mixing and workability, and therefore powders are most often used with plasticizersor superplasticizers.

Regarding workability, it should be noted that a fresh silica fume concrete will havea lower slump than a similar ordinary concrete due to the greater cohesion. When the mixis supplied with energy, as in pumping, vibrating or tamping, the silica fume particles,being spherical, will act as ball bearings and lubricate the mix giving it a greater mobilitythan the similar ordinary concrete. Silica fume concrete is often referred to as beingthixotropic in nature to describe this. Thus when measuring the slump of a silica fumeconcrete it must be remembered that the value will only indicate the consistency of theconcrete and will not relate to its workability. The most favourable test for such concreteis the DIN Standard flow table, or similar, which gives a reaction to an energy input andthus gives a better visual appraisal of the workability of the mix.

In mixes using the slurry material as an addition there can be a slight increase in waterdemand to maintain a given slump. This demand is normally offset by the use of astandard plasticizer. Another way of negating the effect is by reducing the sand content.

In the readymix industry use of the slurry material is nearly always enhanced by anominal dosage of a plasticizer to ensure full dispersion and maintain the w/c ratio. As theconcrete is more cohesive it is less susceptible to segregation, even at very high workabilitiessuch as in flowing or self-compacting concretes. This lack of segregation also makes itideal for incorporation into high-fluidity grout.

The ultrafine nature of the particles will provide a much greater contact surface areabetween the fresh concrete and the substrate or reinforcement and thus will improve thebond between these and the hardened concrete.

Another aspect of this non-segregation, and the filling of the major voids in the freshconcrete, is that a silica fume concrete will produce virtually no bleedwater. The concretemust therefore be cured, in accordance with good site practice, as soon as it has beenplaced, compacted and finished. The lack of bleedwater means that processes such aspowerfloating can be commenced much sooner than with ordinary concretes.

However, this lower slump, lack of bleedwater and ‘gelling’ (stiffening when notagitated) does not indicate a rapid set. Silica fume is a pozzolana and, therefore, requiresthe presence of calcium hydroxide to activate it. The calcium hydroxide is produced bythe cement hydrating and thus the silica fume can only be activated after the cement hasstarted reacting. The setting times for microsilica concretes should be similar to those ofordinary concretes except when specifically designed for such features as ultra-highstrength or low heat.

Mix design criteriaIt is considered by most of the producers of silica fume that the powder forms are bestsuited to specialized production or mixing facilities or precasting and readymix productionusing high power forced action mixers. This is due to the water demand of the powdersand the subsequent difficulty in dispersion in a drier mix.

In most readymix or precast operations it is normal for the slurry product to beintroduced at the same time as the mix water. Although the slurry disperses more readily,care is needed to ensure a uniform quality of mix before allowing the concrete to be used.

Cementitious additions 3/41

There are variations of dosage for given types of application which serve as guidelinesfor initial trial mixes. It should be stressed that, even with the precedent of past work, trialmixes should always be conducted before acceptance of a mix formulation. The values inTable 3.9 are often used as starting points:

Table 3.9

Type of concrete Dosage as % of total cementitious content

Normal 4–7High performance 8–10*High chemical resistance 10–12*Underwater 10–15Pumped 2–5

*Higher values may be used in special circumstances.

The fine filler effect caused by adding a material with a bulk density one fifth of thecement will nearly always require a modification to the mix constituent proportions,particularly the coarse/fine aggregate ratio, to achieve optimum rheology. In nearly allreadymix production a plasticizer – or, more often, a superplasticizer – is used to giveoptimum dispersion, while maintaining the w/c ratio.

In some countries microsilica is added as a large-scale cement replacement in readymixproduction but this practice is decreasing as quality control restrictions become tighter.

Mixing times will need to be adjusted to allow for maximum dispersion of the silicafume. This is most important when using any of the powder forms to prevent agglomerationswithin the mix. The slurry form, when used in readymixed concrete, will not need excessivemixing times since the ‘mass action’ of a truck mixer is known to produce better resultsthan laboratory trials for the same mix.

As with most cementitious materials, silica fume will function more efficiently withsome types of chemical admixtures than with others and trial work is advised. Most ofthe producers of silica fume will have basic mix designs for various situations andaggregate types.

Effects on setting and hardening concreteAs the concrete sets and hardens the pozzolanic action of the silica fume takes over fromthe physical effects. The silica fume reacts with the liberated calcium hydroxide to producecalcium silicate and aluminate hydrates. These both increase the strength and reduce thepermeability by densifying the matrix of the concrete.

Silica fume, having a greater surface area and higher silicon dioxide content, has beenfound to be much more reactive than pfa or ggbs (Regourd, 1983). This increased reactivityappears to increase the rate of hydration of the C3S fraction of the cement in the firstinstance (Andrija, 1986), thus creating more calcium hydroxide, but settles down to morenormal rates beyond two days.

The high reactivity and consumption of calcium hydroxide has prompted questionsrelating to the pH level of the concrete and the corresponding effects on steel passivityand carbonation rate. Studies have shown that the effect on carbonation rate is highlydependent on the quality of the concrete mix produced. Good-quality, well-proportioned,silica fume concrete does not exhibit any greater carbonation than a normal Portland

3/42 Cementitious additions

cement concrete. The reduction of pH in a concrete mix is usually from approximately13.5 to 13.0 and this latter value is well above the level for steel passivity. It has beenestimated that a 25 per cent addition of silica fume (Andrija, 1985) would be required touse up all the calcium hydroxide produced in a concrete, and studies have shown that atthis level the pH still does not drop below 12.0. In normal practice the highest dosageadvised for concrete is 15 per cent and this should have no deleterious effects.

This reduction in alkalinity and the binding of the K+ and Na+ ions in the pore solution(Page and Vennesland, 1983) is one of the ways in which the addition of silica fumedecreases the risk of ASR in concrete (Parker, 1986a).

As the microsilica reacts, and produces the calcium silicate hydrates, the voids andpores within the concrete are filled as the crystals formed bridge the gaps between cementgrains and aggregate particles. Coupling this with the physical filling effect it can be seenthat the matrix of the concrete will be very homogenous and dense, giving improvedstrength and impermeability (Diamond, 1986). It has been found that the relatively poroussection that surrounds the aggregate grains in normal concrete is virtually absent in high-quality silica fume concrete.

The cementing efficiency, a measure of reactivity, of microsilica has been found to bebetween four and five times that of ordinary Portland cement (Page, 1983). This impliesthat large amounts of Portland cement could be replaced by small dosages of silica fumeand a concrete would still achieve the required strength. While this is possible, it is notconsidered to be an ‘ethical’ usage. Reducing the cementitious content to very low levels,though still achieving strength, will have adverse effects on the durability of the concretedespite the benefits imparted by incorporating the silica fume.

Even though the rate of reaction is very high in the initial stages, not all the silica fumeis used up and studies have shown that unreacted material is still present at later stages(Li et al., 1985).

In general, the heat evolution of a silica fume concrete depends on the mix design. Ifa high early strength is needed, addition of silica fume to a high cement content mix willproduce a higher heat of hydration for the initial stages, lessening as the reactivity slows.Additions to ‘normal’ concretes do not usually produce significant changes in heat evolution.

Silica fume concrete is very susceptible to temperature variations during the hardeningprocess. The rate of strength gain can be reduced at temperatures, below the 20°C optimumand accelerated with increased temperatures (Sandvik, 1981). This relates to concrete andnot to some of the specialized repair materials available which can be used at very lowtemperatures.

Blended cement mixesWhile silica fume is compatible with both pulverized fuel ash and ground granulatedblastfurnace slag, it is a pozzolanic material and hence will give differing results dependingon the mix designs used. If present in high proportions the reactivity of pfa will beaffected by the ability of the microsilica to rapidly consume the calcium hydroxide. Thismay provide high early strengths but a reduced rate of long-term gain. The ggbs blendsare less affected because of the latent hydraulic nature of the ggbs. When high replacementlevels of ggbs are used (say 50–70 per cent) silica fume can be added to improve the earlyage strength, as it reacts faster in the first three days, or to improve the consistency of thefresh concrete. High levels of ggbs can cause problems with high water contents leadingto segregation and bleeding, not only on the surface of the concrete but also within the

Cementitious additions 3/43

matrix itself. The silica fume will virtually eliminate this bleeding and hence maintain theintegrity of the concrete. With more normal levels of pfa and ggbs (say 25–40 per cent)silica fume can be added to give enhanced performance. In such cases, where there wouldbe a minor reduction in strength due to using these additions, this is offset by the silicafume and high early, and ultimate, strength can be achieved without an excessive increasein the cost of the concrete. These triple blend cements exploit the beneficial characteristicsof both pozzolanic materials in producing a durable concrete. This type of concrete isbeing specified where concrete structures are expected to last for upwards of 100 yearssuch as the Storebaelt in Denmark and the Tsing Ma bridge in Hong Kong. Here again,caution must be exercised and full trials conducted. While silica fume can enhanceconcretes with high replacement levels of pfa or ggbs, none of them can react properlywithout sufficient cement in the mix to produce calcium hydroxide.

There will always be a point of no return in replacing cement with pozzolanas and thetarget should always be the ultimate quality of the concrete not just the required compressivestrength.

3.6.3 Hardened concrete: mechanical properties

Compressive strengthHigh compressive strength is generally the first property associated with silica fumeconcrete. Many reports are available (Loland, 1983; Loland and Hustad, 1981; Sellevoldand Radjy, 1983) showing that the addition of silica fume to a concrete mix will increasethe strength of that mix by between 30 per cent and 100 per cent dependent on the typeof mix, type of cement, amount of silica fume, use of plasticizers, aggregate types andcuring regimes. In general, for an equal stength, an increase will be seen in the w/c ratiowhile for a given w/c ratio an increased strength will result (Sellevold and Radjy, 1983)(Figure 3.32). Silica fume concrete will show a marked reduction in strength gain whensubjected to early age drying (Johansen, 1981) and this can be as much as 20 per cent.Combinations of silica fume and pfa do not seem as subject to these changes (Maage andHammer, 1985).

With correct design, concrete with ultra-high strength can be produced using normalreadymix facilities. In the USA it is common to use 100–130 MPa concrete in tallbuildings, and 83 MPa silica fume concrete was used to build the 79-storey office blockat 311 South Wacker in Chicago. 100 MPa silica fume concrete was used in the PetronasTowers in Kuala Lumpur. (See also Chapter 1 on High Performance Concrete)

Tensile and flexural strengthThe relationships between tensile, flexural and compressive strengths in silica fumeconcrete are similar to those for ordinary concrete. An increase in the compressive strengthusing silica fume will result in a similar relative increase in the tensile and flexural strength.This plays a strong role when silica fume concrete is used in flooring, bridging or roadwayprojects. The increased tensile strength allows for a possible reduction in slab thickness,maintaining high compressive strengths, thus reducing overall slab weight and cost.

Curing has a large influence on these particular properties and it has been shown thatpoor curing at early ages will significantly reduce the tensile and flexural strength (Johansen,1981).

3/44 Cementitious additions

Brittleness and E modulusThe stronger a concrete is the more brittle it becomes and silica fume concrete is noexception to this rule. However, the strengths concerned are the high to ultra-high values,150–250 MPa and over, and not the 50–100 MPa concrete being used for most projects(Sellevold, 1982a,b; Justesen, 1981).

E modulus does not follow the pattern of tensile strength and only shows slight increasein comparison to compressive strength. Therefore high- and ultra-high strength concretescan be used for tall structures without loss of ductility (Larrard, 1987).

BondingSilica fume concrete has a much finer paste phase and the bond to substrates, old concrete,reinforcement, fibres and aggregates will be improved. Investigation has shown (Carles-Gibergues, 1982) that the aggregate–cement interface is altered when silica fume is present,and pull-out tests (Gjorv et al., 1986; Monteiro et al., 1986) show improved strength.Bonding to fibres is greatly improved (Bache, 1981; Krenchel and Shah, 1985; Ramakrishnanand Srinivasan, 1983). This is particularly beneficial in the steel fibre/silica fumemodified shotcrete which is widely used in Scandinavia. It negates the use of meshreinforcement which would otherwise have to be fixed to the substrate. The high reactivityand extreme cohesiveness of silica fume shotcrete or gunite also reduces the need for anaccelerator. Such use of silica fume has resulted in rebound figures of less than 5 per cent.

ShrinkageShrinkage in cement pastes has been found to be increased when using silica fume and sowhen paste, mortar or grout systems are used, a compensator is usually added.

In concretes the shrinkage is related to the aggregate volume and aggregate quality and

Compressive strength (MPa)

16% MS

8% MS

Reference concrete

0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1

W/C ratio

100

90

80

70

60

50

40

30

20

10

0

Figure 3.32 Strength versus w/c ratio.

Cementitious additions 3/45

many reports are available (Traettenberg and Alstad, 1981; Johansen, 1979), that showthat the addition of silica fume will reduce shrinkage in concrete when close control isexercised over mix design and aggregate selection. The importance of good curing isagain stressed in these papers.

CreepThere is little information about the effect of silica fume on the creep of concrete. Whatis available refers to high-strength, high-dosage mixes and, in general, for such concretesit appears that silica fume concrete exhibits less creep compared to ordinary concrete(Johansen, 1979; Buil and Acker, 1985).

Fire resistanceIt is known that high-strength concretes may explode when exposed to fire. Several tests(Wolsiefer, 1982; Maage and Rueslatten, 1987; Shirley et al.) have shown that undernormal fire conditions silica fume concrete behaves in a similar way to normal concretes.Ultra-high-strength silica fume concrete may be susceptible to this type of failure due toincreased brittleness and consideration should be given to the mix design using low freewater content.

Abrasion and erosionLow w/c ratio high-strength silica fume concrete shows greatly improved resistance toabrasion and erosion and a large amount of silica fume concrete has been produced tospecifically utilize this quality. A large repair project on the Kinzua dam, USA, has beenstudied (Holland, 1983) and results show good performance of the concrete used. Manyhydropower projects in India are utilizing silica fume concrete for this performance.

3.6.4 Hardened concrete: durability-related properties

The use of a silica fume concrete, with its potential for greater strengths, both compressiveand tensile, its more refined pore structure and lower permeability, gives the opportunityof providing a more durable concrete with a longer working lifespan than a conventionalconcrete in the same environment.

PermeabilityThe two main methodologies of measuring permeability are either statically, such asallowing a concrete to dry out and noting the weight loss, or actively, by subjecting thematerial to a liquid or gas under pressure and measuring the depth of penetration. Instudies using drying methods (Sellevold et al., 1982a, b; Sorensen, 1982) the efficiencyfactor for silica fume concrete was between 6 and 8. This indicates that the physical sizeand high reactivity of the silica fume have more influence on permeability than oncompressive strength.

In active testing, the permeability under water pressure, early tests (Sorensen, 1982)showed results ranging from very little permeability to impermeable. Since these tests(carried out in the 1960s) several comprehensive evaluations have been made (Markestad,1977; Hustad and Loland, 1981; Sandvik, 1983) which confirm the previous results.Examinations on mature specimens (Skurdal, 1982) gave reduced permeability and

3/46 Cementitious additions

microscopic studies revealed a very dense microstructure and the virtual absence of theweak layer normally surrounding the aggregate grains.

In all these tests close attention was paid to the curing regimes used and it was foundthat the curing time had a very marked influence on the results. Permeability should notbe confused with porosity, as the pore structure is modified but not decreased (Sellevoldet al., 1982a, b).

The permeability of concretes, particularly to chemicals such as chlorides and sulfates,is a great concern around the world with regard to durability. Research has been ongoingand results are frequently added to the list of references.

Sulfate resistanceIn a major study initiated in Oslo in the first years of testing silica fume concretes variousspecimens were buried in the acidic, sulfate-rich ground in Oslo. The 20-year results areavailable for this trial (Maage, 1984; Fiskaa, 1971) and indicate that the silica fumeconcretes performed as well as those made with sulfate-resisting cement. This is confirmedin the 40-year report and in further laboratory tests (Fiskaa, 1973, Mather, 1980).

When testing in conjunction with ggbs or pfa (Carlsen and Vennesland, 1982), silicafume mixes were found to show greater resistance to sulfate attack than those made withspecial sulfate-resisting cements. This has resulted in silica fume concrete being specifiedin areas such as the Arabian Gulf to combat severe deterioration in the concrete(Rasheeduzzafar et al.).

Such performance of the concrete can be attributed to:

• The refined pore structure and thus the reduced passage of harmful ions (Popovic,1984).

• The increased amount of aluminium incorporated into the microsilica, thus reducingthe amount of alumina available for ettringite formation.

• The lower calcium hydroxide content.

Chloride resistanceChlorides can penetrate concrete from external sources or be present in the mix constituents.The ability of concrete to withstand chloride penetration from, for example, sea water orde-icing salts is important, as is the ability of the mix to bind the aggressive fraction ofthe chlorides present in the pore water. Studies have been made (Page and Vennesland,1983; Mehta, 1981; Page and Havdahl, 1985; Monteiro et al., 1985) which show thevarying effects of the lower permeability and the reduction of pH in the pore water andhow this relates to the presence of chlorides and the state of passivity of embedded steel.

It is considered that the lower the pH, then the lower the threshold limit for depassivationof the steel by chlorides. In studies of the penetration coefficient of chlorides (Fisher,1982) this was shown to be much lower in silica fume concretes, thus negating the effectof the lower threshold value.

In general for equivalent strengths, initiation of chloride attack will be delayed in aconcrete containing silica fume.

Providing sufficient cover to reinforcement is also important since reducing the penetrationrate is ineffective if the cover is also reduced.

In the USA a special rapid test – ASTM 1202 (Christensen et al., 1984) – has beenapplied to compare the chloride penetration of high-strength silica fume concrete with

Cementitious additions 3/47

latex modified and low w/c ratio concrete. The test is based on the resistivity of theconcrete to the passage of an electrical charge. Properly designed, produced and finishedsilica fume concrete normally achieves a very low rating.

CarbonationResults for tests on carbonation rates are somewhat varied and contradictory dependingon the viewpoint taken when analysing the findings. A study (Vennesland and Gjorv,1983) into the effect of microsilica on carbonation and the transport of oxygen showedthat adding up to 20 per cent microsilica caused a slight reduction in these two actions inwater-saturated concrete. In essence, the conclusions shown by those reports available(Johansen, 1981; Vennesland and Gjorv, 1983; Vennesland, 1981) are that for equalstrengths and any concretes below 40 MPa carbonation is higher in silica fume concrete.Concretes above 40 MPa show reductions in carbonation rate and it is only these concretesthat are deemed susceptible to attack and damage if there is reinforcement present.

As silica fume concrete is normally used where the compressive strengths are above40 MPa it is a moot point as to whether carbonation is a serious risk. Correct curingprocedures are essential to ensure optimum performance of the silica fume concrete.

EfflorescenceThese actions occur mainly when one or more surfaces of the concrete is subjected toeither continuous water contact or intermittent wetting and drying. The excess calciumhydroxide is leached through the concrete to the surface where it carbonates, giving awhite powdery deposit. Efflorescence not only reduces the aesthetic quality of structuresbut can also result in increased porosity and permeability and ultimately a weaker andless durable concrete. It has been found in studies (Samuelsson, 1982) that the additionof silica fume will reduce efflorescence due to the refined pore structure and increasedconsumption of the calcium hydroxide. The results indicated that the more efficient thecuring and the longer curing time before exposure, the more resistant the concrete became.

Frost resistanceIt will be appreciated that as the major producers and users in the early years were theScandinavian countries this particular property of silica fume concrete has been wellscrutinized. Investigations have included using silica fume as an addition on its own andwith superplasticizers, varying dosages of air entrainers, different aggregates and variouscuring regimes.

For air-entrained conctete of the required strength it is necessary to achieve the correctamount of air, the right dispersion of the bubbles and a mix stable enough under compaction.

Many different concretes have been compared (Okkenhaug and Gjorv, 1982; Okkenhaug,1983) to determine the effect of silica fume addition. It was found to be difficult to entrainair in a silica fume mix that did not use a plasticizer but that by increasing the dosage ofair entrainer and adding a plasticizer it was easy to achieve the desired levels. There issome speculation as to the reason for the increased dosage of air entrainer with the mostlikely one being that the air and the silica fume compete for the same space in the mix.Once in the mix it was noted that the bubble spacing and stability were greatly improved.The variations of air content for given dosages, as sometimes happens when using pfa,were not noticed in the silica fume concretes.

The use of silica fume with air entrainment is considered to be the best option with the

3/48 Cementitious additions

microsilica maintaining good stability and uniform bubble spacing of the air which givesmaximum frost protection, based around a mix design guideline for a concrete of 30–50 MPa, utilizing 8 per cent microsilica and 5 per cent air entrainment as an optimum. Inall cases the concrete should be cured for the longest allowable time before exposure tothe working environment.

Alkali–silica reactionTo view the effect of silica fume on this form of attack it is necessary to remember thethree main factors that are required for potential reaction.

• A high alkali content in the mix• Reactive aggregates• Available water

Silica fume reacts with the liberated calcium hydroxide to form calcium silicate hydratesand this reduction leads to a lowering of the pH and a lower risk of reaction due to highalkalis. In the formation of the calcium silicate hydrates the K+ and Na+ ions are boundin the matrix and cannot react with any potentially siliceous aggregates.

The minute size and pozzolanic reactivity of the silica fume greatly refines the porestructure of the concrete and reduces the permeability such that less water can enter. Thenormal dosage of microsilica, i.e. 10 per cent by weight of cement, can negate the mainfactors that could lead to alkali–silica reaction and many reports are available (Asgeirssonand Gudmundsson, 1979; Parker, 1986a, b; Perry and Gillott, 1985) that confirm this.

3.6.5 Concluding summary

The inclusion of silica fume in concrete causes significant changes in the structure of thematrix, through both physical action and a pozzolanic reaction, to produce a densified,refined pore system and greater strength. In most cases it is the refinement of the poresystem which reduces penetrability, that has the greater effect on the performance of theconcrete than the increased strength.

Use can be made of these improved qualities in designing concretes to comply withrequirements or greater resistance to certain hostile environments.

Silica fume should be considered as an addition to a mix rather than a replacement forcementitious content and sensible mix design is essential.

Silica fume concrete is susceptible to poor curing and the effects are more pronouncedthan in ordinary concrete. Close attention to curing methods and times is important toensure optimum performance.

Designing a silica fume concrete for specific requirements should always be a matterof consultation between the client, contractor, readymix or precast supplier and the silicafume supplier. Reference should be made to any previous project or use and trial work isessential to ensure correct use of the material and to allow an appreciation of thecharacteristics of this type of concrete.

Cementitious additions 3/49

3.7 Metakaolin

3.7.1 Occurrence and extraction

Kaolin is soft, white clay resulting from the natural decomposition of feldspars and otherclay minerals. It occurs widely in nature. It is used for making porcelain and china, as afiller in the manufacture of paper and textiles and as a medicinal absorbent. Kaolinite isthe principal mineral constituent of kaolin.

In the UK the main sources of kaolin that are commercially exploited are in the south-west of England. The granites that occur in Cornwall originally contained three principalminerals: feldspar, quartz and mica. Over geological time the feldspar decomposed toform kaolin. Kaolin-rich ball clays occur in Devon and Dorset.

Kaolin is extracted from the granite using high-pressure water jets. The kaolin slurryis then concentrated and refined using standard mineral processing techniques. The refinedproduct is dried. The ball clays are treated using standard dry processing techniques.

3.7.2 Metakaolin production

When kaolin is heated to a temperature of 450°C dehydroxylation occurs and the hydratedaluminosilicates are converted to materials consisting predominantly of chemically combinedaluminium, silicon and oxygen. The rate at which water of crystallization is removedincreases with increasing temperature and at 600°C it proceeds to completion (ECCI,1992; Highley, 1984). Metakaolin is formed in kilns when kaolin is heated at a temperaturebetween 700°C and 800°C. The calcined product is cooled rapidly and ground to a finepowder. The metakaolin formed in this way has a highly disorganized structure.

3.7.3 Physical properties

The physical properties of metakaolin depend very much on the quality of the rawmaterial used, the calcination temperature and the finishing processes. The productsavailable in the UK have the typical properties listed below:

Fineness:> 10 microns (mass % max.) 10< 2 microns (mass % min.) 5Surface area (m2/g) 12.0

Moisture:% Water (mass % max.) 0.5Bulk density: (g/cm3) 0.3

3.7.4 Reaction mechanisms

Because of its highly disorganized structure metakaolin reacts very rapidly with thecalcium hydroxide produced during the hydration of Portland cements. The reactivity of

3/50 Cementitious additions

metakaolin may be compared with that of other commonly available industrial pozzolansusing the Chapelle test. In this test, a dilute slurry of the pozzolan is reacted with anexcess of calcium hydroxide at 95°C for 18 hours. After this period the quantity ofcalcium hydroxide consumed in the reaction is calculated. Table 3.10 gives details ofresults of the Chapelle test reported by Largent (1978).

Table 3.10 Reactivity of pozzolans using the Chapelle test

Pozzolan Pozzolan reactivity(mg Ca(OH)2 consumed per gof pozzolan)

Ggbs 40Microsilica 427Pfa 875Metakaolin 1050

More recently Larbi and Bijen (1991) showed that calcium hydroxide was virtuallyeliminated from a cement matrix that contained metakaolin. Jones et al. (1992) confirmedthis, Figure 3.33 summarizes the results of their experiments. In spite of the fact thatcalcium hydroxide levels are reduced significantly, the pH of the pore solution is maintainedabove 12.5 (Asbridge et al., 1992).

Figure 3.33 Effect of metakaolin on the calcium hydroxide content of concrete.

Control10% replacement20% replacement

160

120

80

40

00 10 20 30

Time (days)

Ca(

OH

) 2 c

onte

nt (

arbi

trar

y un

its)

Several workers have studied the products of the reactions between metakaolin andcalcium hydroxide. Turrizani (1964) reported that the cementitious hydrates gehleniteand tobermorite were formed by the reaction. The formation of a number of cementitiouscompounds (calcium silicate hydrates and a rage of calcium aluminosilicate hydrates)and the acceleration of the hydration of C3S have been reported from studies of reactionsbetween metakaolin and calcium hydroxide in the presence of other cement hydrates(Ambroise, 1992; Bredy et al., 1989; Murat, 1983; Pietersen et al., 1993). Dunster et al.,(1993) observed that metakaolin accelerated the early hydration of calcium silicate toproduce extra polymeric silicate gel.

Cementitious additions 3/51

The formation of insoluble, stable cementitious products in place of potentially solublecalcium hydroxide not only reduces the amount available for attack by external agencies(sulfates, acids etc.), it also modifies the pore structure of the paste phase of the concreteleading to reduced porosity and permeability. In his work on the pore solution chemistry,Halliwell (1992) showed that the concentration of alkali metal ions was reduced significantlywhen metakaolin replaced Portland cement in concrete. Table 3.11 shows the results oftests of concrete in which 10 per cent of the cement was replaced by metakaolin.

Table 3.11 Effect of metakaolin content on concentration of alkali metal ions in pore solution

Portland Alkali content Reduction of Alkali metal ioncement (% Na2O equivalent) content of pore water at 1 year (%)

Low alkali 0.32 370.63 46

High alkali 1.15 46

3.7.5 National Standards

No National Standards exist that cover metakaolin as a cementitious addition for concrete.The product Metastar Metakaolin is covered by Agrément Certificate Number 98/3540.This certificate relates to a specific product used at replacement levels up to 20 per centof Portland cement.

3.7.6 The effect of metakaolin on the properties of concrete

Fresh concreteThe particle density of metakaolin (2.4) is lower than that of Portland cement (3.1). Thus,when metakaolin is used as a replacement for cement the volume of cementitious materialis increased. Reducing the sand content of the mix overcomes the effect of the increasedvolume of cementitious powder. Shirvill’s (1992) work indicates that a reduction of sandcontent of 5 per cent is necessary when 10 per cent of cement is replaced by metakaolin.

The cohesiveness of metakaolin concrete is greater than that of plain concrete. Metakaolinconcrete is easier to pump and place generally and it bleeds less than plain concrete. Thegreater volume of cementitious fines results in the production of sharp arrises and high-quality surface finish on cast vertical surfaces.

The greater volume of cementitious material in metakaolin concrete increases its waterdemand. The effect of increasing metakaolin content on the water demand of concrete isshown in Figure 3.34. The increase of water demand is readily offset by the use of asingle dose of a standard low-cost plasticizer (Martin, 1998); this effect is also shown inFigure 3.34

Hardened concreteCompressive strength The effect of metakaolin on the compressive strength of concretehas been widely reported. Some researchers (Larbi and Bijen, 1991; Halliwell, 1992;Saad et al., 1982; Collin-Fevre, 1992) report that metakaolin has no adverse effect oncompressive strength. Other researchers (Andriolo and Sgaraboza, 1986; Marsh, 1992;

3/52 Cementitious additions

Gold and Shirvill, 1992) report significant improvements of strength. This range ofapparently conflicting conclusions reflects differences of factors such as:

• trial mix methodology• approach to mix design• variations of cement composition and• variations in the quality of metakaolin used.

Early work by Gold and Shirvill (1992) showed that the inclusion of metakaolin inconcrete produced very significant improvements in strength. The results of their workindicated that the optimum level of replacement lay somewhere between 5 per cent and10 per cent. Increased replacement, beyond 10 per cent, did not provide further increasesof strength. Sand contents were adjusted for the metakaolin concrete in these trials but noadmixtures were used to offset increased water demands. Figure 3.35 illustrates theresults of these early tests.

Closely monitored plant trials carried out by two RMC Readymix Companies confirmedthe significant increases of strength associated with the inclusion of metakaolin found inlaboratory trials.

The use of metakaolin in high strength concrete has been proposed by a number ofworkers (Marsh, 1992; Balogh, 1995; Calderone et al., 1994). Martin (1995) reported 28-day strengths of 110 N/mm2 for concrete containing metakaolin (10 per cent replacementof Portland cement) and a superplasticizer.

Concrete cast on site is not cured in the ideal conditions to which laboratory curedspecimens are exposed. Hobbs (1996) showed that the compressive strength of plainconcrete cubes and metakaolin concrete cubes was reduced when stored in laboratory air.However, the concrete containing metakaolin was stronger than the plain concrete.

Tensile strength The flexural strength and Young’s modulus of metakaolin concrete aresimilar to those of plain concrete of equivalent 28-day strength. It follows, therefore, thatwhen metakaolin is used to replace a proportion of cement, flexural strength should

Plasticized concretePlain concrete

0 5 10 15 20 25 30Metakaolin content (%)

260

240

220

200

180

Tota

l wat

er c

onte

nt (

L/m

3 )

Figure 3.34 Effect of metakaolin on water demand.

Cementitious additions 3/53

increase. A study at Dundee University, reported by Imerys (2000), confirmed this to bethe case. Table 3.12 summarizes the results of the experiments.

5% replacementMean 10% and 20% replacementControl

200 300 400 500 600

Cementitious material content (kg/m3)

70

65

60

55

50

45

40

28-d

ay c

ompr

essi

ve s

tren

gth

(N/m

m2 )

Figure 3.35 Effect of metakaolin on compressive strength of concrete.

Table 3.12 Strength properties of concrete with a range of metakaolin contents

Mix Cube strength Flexural strength Young’s(MPa) (MPa) modulus (Gpa)

100% PC 41.0 5.0 30.090% PC, 10% 47.0 Not measured 33.0Mk80% PC, 20% 50.0 5.3 33.5Mk

The development of tensile strength at very early ages has been studied (Hobbs, 1996).The trials showed that metakaolin significantly accelerated the development of earlytensile strength.

Creep strain Creep strain has been measured for plain concrete and concrete containingmetakaolin (Imerys, 2000). For cubes cured in water and then loaded to 40 per cent oftheir 28-day strength for 90 days, creep strain was unaffected by the inclusion of metakaolin.

Drying shrinkage The incorporation of metakaolin in concrete produces a cementitiousmatrix of low porosity and permeability. Water loss on drying is therefore reduced anddrying shrinkage is correspondingly less. Zhang and Malhotra (1995) showed that concretecontaining 10 per cent metakaolin had a lower drying shrinkage than that of plain concreteand concrete containing microsilica. Using similar test procedures Caldarone et al. (1994)showed that the drying shrinkage of metakaolin concrete was lower than that of plainconcrete but similar to that of concrete containing microsilica.

More recently, work carried out at Dundee University (Imerys, 2000), showed thatmetakaolin levels up to 25 per cent did not significantly increase drying shrinkage.

3/54 Cementitious additions

3.7.7 The durability of metakaolin concrete

Durability of concrete may be defined as its ability to resist the action of weathering,chemical attack, abrasion or any other process of deterioration. In recent years thisproperty has gained a much higher profile with engineers and specifying authorities.

With the exception of mechanical damage, many of the adverse influences on durabilityinvolve the transport of fluids through the concrete. Therefore, in general terms, high-strength low-permeability concrete should provide adequate durability. Additionalconsiderations may be necessary in some circumstances, e.g. alkali–silica reaction. Concreteis less able to withstand some external chemical effects, e.g. acid attack. However, it ispossible to provide concrete that, though affected by acids, has a significantly prolongedservice life.

Resistance to freezing and thawingThe resistance of concrete to cyclic freezing and thawing depends upon a number ofproperties (e.g. strength of the hardened paste, extensibility and creep) but the mainfactors are the degree of saturation and the pore system of the hardened paste (Neville,1995). The use of metakaolin in concrete significantly alters the pore system and permeabilityof the paste. It follows, therefore, that well designed metakaolin concrete will haveincreased resistance to the damaging effects of freezing and thawing. Zhang and Malhotra(1995) report that metakaolin concrete showed excellent performance after 300 cycles offreezing and thawing in the ASTM C666 test. The results of experiments by Calderone etal. (1994) confirmed that metakaolin concrete exhibited high freeze–thaw durability.

Sulfate resistanceSingh and Osborne (1994) showed that mortar prisms containing 15 per cent metakaolinshowed good sulfate resistance when stored in 4.4 per cent sodium sulfate solution at20°C. Khatib and Wild (1998) confirmed these findings. When stored at 5°C in bothsodium sulfate and magnesium sulfate, sulfate resistance was further improved. Giventhe concerns about the formation of thaumasite at low temperatures and the subsequentdeterioration of the concrete, the results of tests at 5°C are of particular interest.

Resistance to the penetration of chloride ionsThe ASTM test C1202:1944 has been used to assess the ability of metakaolin concreteto resist the penetration of chloride ions. Calderone et al. (1994) and Zhang and Malhotra(1995) report that the resistance of metakaolin concrete to the penetration of chloride ionswas significantly higher than that of plain concrete of otherwise similar composition.Ryle’s work (1994) compared the results of tests using the ASTM procedure with otherchloride and oxygen diffusion test methods and confirmed that concrete of very lowpermeability was produced when 15 per cent of the cement was replaced with metakaolin.

Table 3.13 summarizes results of tests using the ASTM procedure extracted from theliterature.

Page and Coleman (1994) and Larbi and Bijen (1992) showed that the use of metakaolinas a partial replacement for cement (15 per cent and 20 per cent) significantly affectedthe transport processes in mortar and concrete. They showed that the partial replacementof cement with metakaolin led to a reduction of the rate of diffusion of chloride ionsinto concrete and mortar by an order of magnitude. Table 13.14 shows the effect of

Cementitious additions 3/55

the use of metakaolin on chloride ion diffusion coefficients for mortar and paste (ECCI,1993).

Resistance to acidic conditionsMetakaolin reacts very rapidly with the calcium hydroxide produced during hydrationconverting it to a variety of insoluble, stable cementitious products. In addition, byreducing the permeability of the paste and reducing the thickness of the interface betweenthe paste and the aggregates (Larbi and Bijen, 1991, 1992; Mass, 1996) metakaolin helpsto prevent the ingress of aggressive substances. It follows, therefore, that concrete containingmetakaolin as a partial replacement for cement will prolong the service life of concretesubjected to attack by acidic substances.

Collin-Fevre (1992) showed that concrete containing 10 per cent metakolin as a partialreplacement for cement was considerably more resistant to attack by organic and mineralacids than plain concrete of otherwise similar composition.

Dutrel and Estoup (1986) reported that concrete containing metakaolin (10 per centreplacement of Portland cement) was more resistant to the effects of lactic, formic, citricand humic acids and to mineral acids of pH 3.5.

Experiments by Martin (1997) showed that concrete containing metakaolin was moreresistant than plain concrete to the acids that result from the storage of silage. To complementhis laboratory work, a working silage bay was constructed using plain concrete andconcrete containing metakaolin. After some 10 years in use no adverse effects are apparenton the surface of the area constructed with metakaolin concrete. Attack by silage acidshad removed some of the surface of the plain concrete after 5 years of use.

Table 3.13 Results of chloride ion penetration tests (ASTM 1202:1994)

Research Mix Cementitious material Results of ASTM test (coulombs)reference content ———————————————

(kg/m3) 28 days 56 days 90 days

Calderone Control 385 – 4832 –15% Mk 385 – 754 –

Martin Control 385 3175 – 187515% Mk 385 390 – 300

Ryle 15% Mk 400 625 395 310

Table 3.14 The effect of metakaolin on the chloride ion diffusion coefficient

Sample type W/C Ratio % PC Replaced with Chloride ion diffusion coefficientmetakaolin (cm2 s–1) × 10–9

Paste 0.5 0 900.5 15 9

Mortar 0.6 0 2200.6 15 18

Mortar 0.7 0 6800.7 15 8

Mortar – 0 597– 20 10

3/56 Cementitious additions

Alkali–silica reactionThe suppression of alkali–silica reaction using metakaolin in concrete containing reactivecombinations of aggregates is well established by laboratory experiment (Jones, et al.,1992; Martin, 1993; Kostuch, 1993). Sibbick and Nixon (2000) showed that the inclusionof metakaolin in larger concrete castings stored on outdoor exposure sites preventedexpansion due to alkali–silica reaction.

The earliest large-scale use of metakaolin in concrete dates back to the 1960s. Some300 000 tonnes of locally calcined kaolin was blended with Portland cement for theconstruction of a series of four dams in the Amazon basin. The only aggregates availablewere alkali reactive. Andriolo and Sgaraboza (1986) claimed that the use of metakaolinprevented expansion due to alkali–silica reaction.

Andriolo and Sgaraboza (1986) reported additional beneficial effects resulting fromthe use of metakaolin. They claimed.

• substantially increased compressive strength of concrete, permitting a reduction in theamount of binder used

• reduced bleeding of the concrete• substantially reduced temperature rise in mass concrete.

These benefits were of particular significance in view of the large volumes of concreteused in the construction of the dams.

Jones et al. (1992) showed that metakaolin reduced the calcium hydroxide content ofthe concrete to such a low level that the production of a swelling gel was not possible.

The effect of alkalis from external sources on the alkali–silica reaction has also beendemonstrated (Walters and Jones, 1991; Xu and Chen, 1986). In Figure 3.36 the effect ofimmersing concrete prisms in sodium chloride solution is shown. Prior to immersion theprisms had been stored at a temperature of 38°C and a relative humidity close to 100 percent to accelerate any potential reactions. By (about) 20 months the reaction in theexpansive control mix had reached equilibrium. Immersion in sodium chloride solutiondisturbed this equilibrium and presented the control mix and the mix containing themetakaolin with additional alkalis. This caused the control mix prisms to expand rapidlyagain. The prisms containing the metakaolin concrete remained unaffected.

Figure 3.36 The effect of metakaolin on expansion due to ASR and immersion in NaCl.

Control15% metakaolin

Transferred tosaturated NaCl

0 20 40 60Time (months)

2

1.5

1

0.5

0

Exp

ansi

on (

%)

Cementitious additions 3/57

3.7.8 Compatibility with blended cements

A large proportion of readymixed concrete manufactured in the UK contains cementreplacement materials such as ground granulated blast-furnace slag (ggbs) and pulverizedfuel ash (pfa). These materials are industrial pozzolans and are activated in much thesame way as metakaolin.

Asbridge et al. (1994) showed that when used in conjunction with ggbs, metakaolinincreased the strength of concrete at 96 days. The pore volume of the paste was lowerthan that of the corresponding control mixes containing ggbs. They considered that thesechanges would further improve the durability of concrete. Unpublished work by RMCReadymix (1999) shows that higher strengths and improved durability were achievedwhen metakaolin was used as a partial replacement for cement containing 50 per centggbs.

3.7.9 Efflorescence

Efflorescence is a white deposit that appears on concrete surfaces. The deposit does notaffect the durability of the concrete but it detracts from the aesthetic properties of plainand coloured concrete.

There are two forms of efflorescence: primary and secondary. Primary efflorescenceoccurs early in the life of a structure, during the process of curing. Lime-rich water fromwithin the concrete matrix migrates to the surface and evaporates, depositing soluble saltsat the surface. The lime (Ca(OH)2) reacts with atmospheric carbon dioxide to forminsoluble calcium carbonate. Secondary efflorescence occurs when hardened concrete iswetted and water penetrates the surface to dissolve some of the lime remaining in theconcrete. During subsequnt drying the salt-laden solution moves to the surface depositingthe salts at the surface.

Metakaolin is thought to control the formation of efflorescence by:

• reducing the alkali content of the concrete, so CO2 is absorbed less rapidly• removing a proportion of the Ca(OH)2

• refining the pore structure so the absorption of water and the diffusion of salt-ladenwater to the surface is reduced.

3.7.10 Summary

When kaolin is heated to a temperature in the range 700–800°C it is converted to metakaolinwith a highly disorganized structure. Metakaolin is a very reactive pozzolan. In thepresence of water, it reacts with calcium hydroxide to produce stable, insoluble cementitioushydrates. This pozzolanic reaction reduces the permeability and porosity of cement pastemaking it stronger and significantly more durable. The use of metakaolin as a partialreplacement for cement in suitably designed concrete mixes has been shown to:

• improve cohesion and reduce bleeding of fresh concrete• increase compressive strength• reduce drying shrinkage

3/58 Cementitious additions

• improve freeze–thaw resistance• improve sulfate resistance• increase resistance to the penetration of chloride ions• improve acid resistance• eliminate alkali–silica reaction.

3.8 Limestone

3.8.1 Limestone filler

Limestone fillers are being used increasingly in cements and, more recently, as a mixeraddition. EN197-1 (BS EN197-1, 2000) contains two classes of Limestone Portlandcement, CEM II/A-L (or LL) and II/B-L (or LL). The former contains between 6 per centand 20 per cent limestone and the latter 21–35 per cent limestone. This material isinterground with the Portland cement clinker, which normally produces 32.5 grade cement.The requirements for the limestone are:

• The CaCO3 content ≥ 75 per cent• The clay content, as determined by the methylene blue test shall not exceed 1.20 g/

100 g.• The total Organic Carbon (TOC) shall not exceed 0.20 per cent for LL limestone or

0.50 per cent for L limestone.

The limestone filler was considered by many as inert filler, but it has been graduallyaccepted as contributing to the hydration process by the formation of calcium mono-carboaluminates (C3A · CaCO3 · 11H2O).

In addition to factory-made Portland limestone cements, limestone fines complyingwith BS 7979 (2001) may also be added at the mixer to produce an equivalent combination.The specification for the filler is very similar to that for cement. Limestone fillers mayalso be used in special concrete, such as self-compacting concrete, to aid cohesion in theplastic concrete.

Classification of additions to BS EN206-1Additions, that is, materials like fly ash, silica fume, limestone fillers, etc., are classifiedwithin EN206-1 (BSI, 2000) in two ways. They are defined within the standard as being‘finely divided materials used in concrete in order to improve certain properties or toachieve special properties. This standard deals with two types of additions:

• Nearly inert additions (Type I).• Pozzolanic or latent hydraulic additions (Type II).’

The practicality of this definition is that some additions are treated as aggregate and somemay be counted towards the cement content, either fully or partially. The k-value conceptis applied to fly ash and silica fume within EN206-1, which are defined as Type IIadditions.

Smith (1967) developed a method based on applying a cementing efficiency factorknown as the k factor or k-value. The mix design was adjusted as in

Cementitious additions 3/59

W/Cf = W/(C + k · F )

where W/C in the equation for plain Portland cement concrete is replaced by the adjustedW/Cf ratioW = weight of waterC = weight of Portland cementCf = equivalent weight of Portland cementk = cement efficiency factor for fly ashF = weight of fly ash

k-values can be created/applied for many purposes, e.g. equal 28-day strength, equalchloride diffusion, equal durability, etc. A calculated k-value will change depending onthe Portland cement source, the curing temperature and conditions, the fly ash source, etc.The technique has been corrupted in EN206-1 that gives k-values for fly ash of 0.40 foruse in combination with CEM I 42.5N or 0.20 for CEM I 32.5. These k-values are usedto adjust minimum cement contents and maximum water cement ratios.

In real concretes it will be found the k-value for a fly ash, for example, will varyconsiderably from zero to greater than unity, depending on the materials, the temperature,the time, etc. Consequently, the k-value is one of the most variable constants known to theconcrete industry.

The suitability of additions, whether they are considered to be Type I or II, the k-valueor equivalent performance concepts and other factors are allowed on a National basis,providing their suitability has been established. The UK complementary standard thatgives the relevant information is BS 8500 (BSI).

References

ACI (1994) Manual of Concrete Practice, Fly ash, 226.3R.Aitcin, P.C. and Regourd, M. (1985) The use of condensed silica fume to control alkali silica

reaction – a field case study. Cement and Concrete Research, 15, 711–719.Alasali, M.M. and Malhotra, V.M. (1991) Role of concrete incorporating high volumes of fly ash

in controlling expansion due to alkali–aggregate reaction. ACI Materials Journal, 88, No. 2,159–163.

Ambroise, J., Martin-Calle, S. and Pera, J. (1992) Properties of metakaolin blended cements. 4thInt. Conf. Flyash, silica fume, slag and natural pozzolans in concrete, Volume 1, Turkey.

Andrija D. (1986) 8th International Congress on Chemistry of Cement, Rio de Janeiro, Brazil.Volume 4, pp. 279–285.

Andriolo, F.R. and Sgaraboza, B.C. (1986) The use of pozzolan from calcined clays in preventingexcessive expansion due to alkali–silica reaction in some Brazilian dams. Proc. 7th Int. Conf. onAAR. Ottawa.

Asbridge, A.H. et al. (1994) Ternary blended concretes-OPC/ggbs/metakaolin. Proc. Int. Sym.Concrete across borders. Denmark.

Asbridge, A.H., Jones, T.R. and Osborne, G.J. (1996) High performance metakaolin concrete:Results of large-scale trials in aggressive environments. Proc. Int. Conf. Concrete in the serviceof mankind, Dundee.

Asgeirsson, H. and Gudmundsson G. (1979) Pozzolanic activity of silica dust. Cement and ConcreteResearch, 9, 249–252.

Bache, H.H. (1981) Densified cement/ultrafine particle-based materials. 2nd International Conferenceon Superplasticisers in Concrete, Ottawa.

3/60 Cementitious additions

Balogh, A. (1995) High reactivity metakaolin. Concrete Construction.Bamforth, P.B. (1984) Heat of hydration of fly ash concrete and its effect on strength development.

Ashtech ’84 Conf. London, 287–294.Bamforth, P.B. (1993) Concrete classification for R.C. structures exposed to marine and other salt

laden environments. Proceedings, Conference on Structural Faults & Repairs. Edinburgh.Bamforth, P.B. Taywood Engineering, England, private communication.Bensted, J. (1988) Thaumasite – a deterioration product of hardened cement structures. II Cemento,

3–10.Bensted, J. and Barnes, P. (2001) Structure and performance of cements, 2nd edn, E & FN Spon,

London.Berry, E.E. and Malhotra, V.M. Fly ash in concrete. CANMET, SP85-3.Bredy, P., Chabannet, M. and Pera, J. (1989) Microstructure and porosity of metakaolin blended

cements. Proc. Mats. Res. Soc. Sym. Boston.Brown, J.H. (1980) The effect of two different pulverised fuel ashes upon the workability and

strength of concrete. C&CA Technical Report No. 536, June.Browne, R.D. (1984) Ash concrete – its engineering performance. Ash Tech ’84, London, 295–301.BS 146. Specification for Portland-blastfurnace cement.BS 4246. Specification for low heat Portland-blastfurnace cement.BS 4550. Methods of testing cement.BS 6699. Specification for ground granulated blastfurnace slag for use with Portland cement.BS 8110. Structural use of concrete.BS EN197-1 (2000) Cement – Part 1: Composition, specifications and conformity criteria for

common cements, BSI, London.BS EN206-1 (2000) Concrete – Part 1: Specification, performance, production and conformity,

BSI, London.BS 7979 (2001) Specification for limestone fines for use with Portland cement, BSI, London.BS 8110. Structural use of concrete, BSI, London.BS 8500-1. Concrete – complementary British standard to BS EN206-1, Part 1: Method of specifying

and guidance for the specifier, BSI, London.BS 8500-2. Concrete – complementary British standard to BS EN206-1, Part 2: Specification for

constituent materials and concrete, BSI, London.BS 5328: Part 2 (1997) Methods for specifying concrete, amendment 10365, May, 1999.Buil, M. and Acker, P. (1985) Creep of silica fume concrete. Cement and Concrete Research, 15,

463–466.Building Research Establishment (1991) Sulfate and acid resistance of concrete in the ground. BRE

Digest 363, July.Building Research Establishment (1999) Alkali–silica reaction in concrete. Digest No 330. BRE,

Garston.Building Research Establishment (2001) Concrete in aggressive ground. Special Digest No. 1

2001, BRE, Garston.Burton, M.W. (1980) The Sulphate Resistance of concretes made with ordinary Portland cement,

sulphate resisting cement and ordinary Portland cement + pozzolan. Kirton Concrete Services,Humberside.

Cabrera, J.G. and Atis, C.D. (1998) Design and properties of high volume fly ash performanceconcrete.

Cabrera, J.G., Braim, M. and Rawcliffe, J. (1984) The use of pulverised fuel ash for constructionof structural fill. Ash Tech, London.

Cabrera, J.G. and Plowman, C. (1987) Hydration and microstructure of high fly ash content concrete.Conference on concrete dams, London.

Cabrera, J.G. and Woolley, G.R. (1996) Properties of sprayed concrete containing ordinary Portlandcement or fly ash Portland cement. Proc. ACI/SCA International Conference, Edinburgh.

Caldarone, M.A. et al. (1994) High reactivity metakaolin: A new generation mineral admixture.Concrete International.

Cementitious additions 3/61

Carette, G.G. and Malhotra, V.M. (1983) Mechanical properties, durability and drying shrinkage ofPortland cement concrete incorporating silica fume. Cement, Concrete and Aggregates, 5, No.1, 3–13.

Carles-Gibergues, A. et al. (1982) Contact zone between cement paste and aggregate. In Bartos, P.(ed.), Proc. International Conference on Bond in Concrete. Applied Science Publishers, London,24–33.

Carlsen, R. and Vennesland, O. (1982) Sementers sulfat- og sjovannsbestandighet. FCB/SINTEF,Norwegian Institute of Technology, Trondheim, Report STF65 F82010 (in Norwegian).

Christensen, D.W. et al. (1984) Rockbond: a new microsilica concrete bridge deck overlay material.Proc. International Bridge Conference, Pittsburgh, 151–160.

Christensen, P. (1982) Afprovningaf kvalitetskarakterisering af beton tilsat Silica. TeknologiskIsntitut. Byggeteknik, Copenhagen (in Danish).

Clear, C.A. and Harrison, T.A. Concrete pressure on formwork. Report 108, CIRIA, London.Collin-Fevre, I. (1992) Use of metakaolin in the manufacture of concrete products. CIB, Montreal.Concrete Society (1991) The use of GGBS and fly ash in concrete, Technical report 40, Crowthorne.Concrete Society (1998) Core Project, project data for a potential revision of Technical Report No.

11.Concrete Society (1999) Alkali silica reaction: minimising the risk of damage. Technical report No.

30, 3rd edition.Concrete Society. Microsilica in Concrete. Technical Report 4/1993.Davis, R.E., Carlston, R.W., Kelly, J.W. and Davis, H.E. (1937) Properties of cements and concretes

containing fly ash. ACI Journal, 33, 577–612.Davis, R.E., Kelly, J.W., Troxell, G.E. and Davis, H.E. (1935) Proportions of mortars and concretes

containing Portland-pozzolan cements. ACI Journal, 32, 80–114.Dewar, J.D. (1986) The particle structure of fresh concrete – a new solution to an old question. Sir

Frederick Lea Memorial Lecture, Institute of Concrete Technology Annual Symposium.Dhir, R.K. (1986) Pulverised fuel ash. CEGB Ash-Tech 86 conference proceedings.Dhir, R.K., Munday, J.G.L. and Ho, N.Y. (1987) Fly ash in concrete: freeze–thaw durability. Draft

report, University of Dundee.Dhir, R.K., Munday, J.G.L. and Ong, L.T. (1981) Strength variability of OPC/fly ash concrete.

Concrete, June.Dhir, R.K., Munday, J.G.L. and Ong, L.T. (1986) Investigations of the engineering properties of

OPC/Pulverised fuel ash concrete – deformation properties. The Structural Engineer, 64B, No.2, 36–42.

Diamond, S. (1986) 8th International Congress on Chemistry of Cement, Rio de Janeiro Brazil,Vol. 1, 122–147.

Dunstan, M.R.H. (1981) Rolled Concrete for Dams. CIRIA Technical note No. 106, London.Dunster, A.M. et al. (1993) Journal Mat. Sci.Dutrel, F. and Estoup, J.M. (1986) Effect of the addition of silica fume or pozzolanic fines on the

durability of concrete products. Madrid.ECC International. Calcined clay. Internal data sheet.ECCI (1993) New pozzolanic materials for the concrete industry. UK.European Standard ENV 197-1 (1992) Cement – composition, specifications and conformity criteria:

Common cements.FCB/SINTEF (1977) Norwegian Institute of Technology, Trondheim, Report STF65 A77027.FCB/SINTEF (1984) Norwegian Institute of Technology, Trondheim, Report STF65 A84019.Feldman, R.F. and Huang Cheng-Yi. (1984) Microstructural properties of blended cement mortars

and their relation to durability. RILEM Seminar on Durability of Concrete Structures underNormal outdoor Exposure, Hanover.

FIP (1988) State of Art Report, Condensed silica fume in conctete, Thomas Telford, London.Fisher, K.P. et al. (1982) Corrosion of steel in concrete: some fundamental aspects of concrete with

added silica. Norwegian Geotechnical Institute, Oslo, Report no. 51304-06.

3/62 Cementitious additions

Fiskaa, O. et al. (1971) Betong i Alunskifier. Norwegian Geotechnical Institute, Oslo, Publication86. (in Norwegian).

Fiskaa, O.M. (1973) Betong i Alkunskifier. Norwegian Geotechnical Institute, Oslo, Publication101 (in Norwegian).

Fournier, B. and Malhotra, V.M. (1997) CANMET investigations on the effectiveness of fly ash inreducing expansion due to alkali aggregate reaction (ASR), ACAA 12th International Symposium.

Fulton, A.A. and Marshall, W.T. (1956) The use of fly ash and similar materials in concrete. Proc.Inst. Civ. Engrs. Part 1, Vol. 5, 714–730.

Gifford, P.M. and Ward, M.A. (1982) Results of laboratory test on lean mass concrete utilising flyash to a high level of cement replacement. Proc. International Symposium, Leeds, 221–229.

Gjorv, O.E. et al. (1986) Effect of condensed silica fume on the steel–concrete bond. NorwegianInstitute of Technology, Trondheim, Report BML 86.201.

Glasser, F.P. and Marr, J. (1984) The effect of mineral additives on the composition of cement porefluids. Proceedings No. 35, The Chemistry and Chemically Related Properties of Cement,British Ceramic Soc., 419–429.

Gold, S.J. and Shirvill, A.J. (1992) Effects of metakaolin on concrete strength. Laboratory Report22 RMC Readymix Limited.

Grube, H. (1985) Influence of concrete materials mix design and construction techniques onpermeability. Concrete Society Conference, Permeability of Concrete, London.

Halliwell, M.A. (1992) Preliminary assessment of the performance of Portland cement concretecontaining metakaolin. BRE Client Report TCR48/92. BRE.

Hewlett, P. (ed.), Lea’s ‘The Chemistry of Cement and Concrete’, 4th edn. (Chapter 12: ‘Microsilicaas an addition’).

Highley, D.E. (1984) China clay. Mineral Dossier No. 26. HMSO, London.Hobbs, D.W. (1982) Influence of pulverised-fuel ash and granulated blastfurnace slag upon expansion

caused by the alkali-silica reaction. Concrete Research, 34, No. 119, 83–94.Hobbs, M. (1996) Project Report, University of Surrey.Holland, T.C. (1983) Abrasion-erosion evaluation of concrete admixtures for stilling basin repairs,

Kinzua Dam, Pennsylvania. US Army Engineer Waterways Experiment Station, StructuresLaboratory, Vicksburg, Miscellaneous Paper SL-83-16.

Hooton, R.D. (1987) Some aspects of durability with condensed silica fume in pastes, mortars andconcretes. International Workshop on Condensed Silica Fume in Concrete, Montreal.

Huang Cheng-Yi and Feldman R.F. (1985) Dependence of frost resistance on the pore structure ofmortar containing silica fume. ACI Journal, Sept.–Oct.

Hustad. T. and Loland, K.E. (1981) Report 4: Permeability. FCB/SINTEF, Norwegian Institute ofTechnology, Trondheim, Report STF65 A81031.

Imerys, (2000) Effect of Metastar on the mechanical properties of concrete. Information sheetPMA30/MK.

Johansen, R. (1979) Silicastov i fabrikksbetong. Langtidseffekter. FCB/SINTEF, Norwegian Instituteof Technology, Trondheim, Report STF65 F79019.

Johansen, R. (1981) Report 6: long term effects. FCB/SINTEF, Norwegian Institute of Technology,Trondheim, Report STF65 A81031.

Jones, T.R., Walters, G.V. and Kostuch, J.A. (1992) Role of metakaolin in suppressing asr inconcrete containing reactive aggregate and exposed to NaCl solution. Proc. 9th Int. Conf. onAAR. London.

Jorgen, J.A.H.R. (1980) Possible health hazards from different types of amorphous silicas – suggestedthreshold limit values. Institute of Occupational Health, Oslo. HD806/79. Revised.

Joshi, R.C. and Lohtia, R.P. (1997) Advances in Concrete Technology, Volume 2, Fly Ash inConcrete, Gordon and Breach Science Publishers, Now York.

Justesen, C.F. (1981) Performance of dense injection grout for prestressing tendons. TechnicalAdvisory Service, Aalbory Portland, Aalborg, Denmark.

Keck, R.H. and Riggs, E.H. (1997) Specifying fly ash for durable concrete. Concrete International,April.

Cementitious additions 3/63

Kostuch, J.A. et al. (1993) High performance concretes incorporating metakaolin – a review. Proc.Int. Conf. Concrete 2000, Dundee.

Krenchel, H. and Shah, S. (1985) Applications of polypropylene fibres in Scandinavia. ConcreteInternational; Design and Construction, 2, No. 3, 32–34.

Kumar, A., Roy, D.M. and Higgins, D.D. (1987) Diffusion through concrete. Concrete, 21, No 1,31–34, Jan.

Larbi, J.A. and Bijen, J.M. (1991) The role of silica fume and metakaolinite in the Portland cementpaste–aggregate interfacial zone in relation to the strength of mortars. PhD thesis, Delft University.

Larbi, J.A. and Bijen, J.M. (1992) Influence of pozzolans on the Portland cement aggregate interfacein relation to diffusion of ions and water absorption of concrete. Delft University.

Largent, R. (1978) Research Bulletin. Liaisons LCPC, France.Larrard, F. et al. (1987) (Fracture toughness of high strength concrete. Proc. Utilisation of High

Strength Concrete, Stavanger, 215–223.Li, S., Roy, D.M. and Kumar, A. (1985) Quantitative determination of pozzolanas in hydrated

systems of cement or Ca(OH)2 with fly ash or silica fume. Cement and Concrete Research, 15,1079–1086.

Loland, K.E. and Hustad, T. (1981) Report 2: Mechanical properties. FCB/SINTEF, NorwegianInstitute of Technology, Trondheim, Report STF65 A81031.

Loland, K.E. (1983) Fasthets-og deformasjonsegenskaper i herdnet tilstand – herdebetingelser.Seminar Bruk av silika i betong, Norsk Sivilingeniorers Forening, Oslo (in Norwegian).

Maage, M. Effect of microsilica on the durability of concrete structures.Maage, M. and Hammer, T.A. (1985) Modifisert Portlandsement. Delrapport 3. Fasthetsutvikling

og E-modul. FCB/SINTEF, Norwegian Institute of Technology, Trondheim, Report STF65A85041 (in Norwegian).

Maage, M. and Rueslatten, H. (1987) Trykkfasthet og blaeredannelse pa brannpakjent hoyfastbetong.FCB/SINTEF, Norwegian Institute of Technology, Trondheim, Report STF65 A87006 (inNorwegian).

Malhotra, V.M. et al. (1987) Mechanical properties and freezing and thawing resistance of highstrength concrete incorporating silica fume. International Workshop on Condensed Silica Fumein Concrete, Montreal.

Malhotra, V.M. and Mehta, P.K. (1996) Advances in Concrete Technology, Volume 1, Pozzolanicand Cementitious Materials, Gordon and Breach Science Publishers, New York.

Markestad, S.A. (1977) An investigation of concrete in regard to permeability problems and factorsinfluencing the results of permeability tests.

Marsh, D. (1992) An alternative to silica fume? Concrete Products.Martin, S.J. (1993a) Suppression of ASR using metakaolin, ggbs, pfa, and microsilica. Laboratory

Report 24, RMC Readymix Limited.Martin, S.J. (1993b) Suppression of ASR using metakaolin. Laboratory Report 33, RMC Readymix

Limited.Martin, S.J. (1995) The use of metakaolin in high strength concrete. Laboratory Report 78, RMC

Readymix Limited.Martin, S.J. (1998) Water demand and other characteristics of metakaolin concrete. Laboratory

Report 135, RMC Readymix limited.Martin, S.J. (1997) Metakaolin and its contribution to the acid resistance of concrete. Proc. Int.

Sym. Concrete for a sustainable agriculture, Stavanger.Mass, J.C. (ed.) (1996) Interfacial transition zone in concrete. Rilem Report 11.Mather, K. (1980) Factors affecting the sulphate resistance of mortars. Proc. 7th International

Conference on Chemistry of Cements, 4, 580–585.Matthews, J.D. and Gutt, W.H. (1978) Studies of fly ash as a cementitious material. Conference on

Ash Technology and Marketing, London, October.McMillan, F.R. and Powers, T.C. (1934) A method of evaluating admixtures, Proceedings American

Concrete Institute, Vol. 30, 325–344, March–April.

3/64 Cementitious additions

Mehta, P.K. (1981) Sulphate resistance of blended Portland cements containing pozzolans andgranulated blastfurnace slag. Proc. 5th International Symposium on Concrete Technology, Monterey.

Monk, M.G. (1983), Portland-fly ash cement: A comparison between intergrinding and blending.Concrete Research, 35, 124, September, 131–141.

Monteiro, P.J.M. et al. (1985) Microstructure of the steel–cement paste interface in the presence ofchloride. Cement and Concrete Research, 15, 781–784.

Monteiro, P.J. et al. (1986) Effect of condensed silica fume on the steel–cement paste transitionzone. Norwegian Institute of Technology, Trondheim, Report BML 86.205.

Murat, M. (1983) Hydration reaction and hardening of calcined clay and related minerals. Preliminaryinvestigation of metakaolin. Cement and Concrete Research, 13.

Neville, A.M. (1995) Properties of Concrete (4th edn), Longman, London.Neville, A.M. and Brooks, J.J. (1975) Time-dependent behaviour of Cemsave concrete. Concrete,

9, No. 3, 36–39.Nustone Environmental Trust (2000) The effect of the fineness of PFA (fly ash) on the consistence

and strength properties of standard mortar. Project report, March. Read in conjunction withEN450 Fly Ash BS 3892 Part 1 Testing Program, UKQAA, Analysis of results, March 2000.

Okkenhaug, K. (1983) Silikastovets innvirkning pa luftens stabilitet i betong med L-stoff og medL-stoff i kombinasjon med P-stoff. Swedish Cement and Concrete Research Institute at theInstitute of Technology, Stockholm CBI report 2:83, 101–105 (in Norwegian).

Okkenhaug, K. and Gjorv, O.E. (1982) Influence of condensed silica fume on the air-void systemin concrete. FCB/SINTEF, Norwegian Institute of Technology, Trondheim, Report STF65 A82044.

Osborne, G.J. (1986) Carbonation of blastfurnace slag cement concretes. Durability of BuildingMaterials, 4, 81–96.

Owens, P.L. (1979) Adapted and redrawn from ‘Fly ash and its usage in concrete’. Concrete, July.Page, C.L. (1983) Influence of microsilica on compressive strength of concrete made from British

cement and aggregates. Elkem Materials Ltd.Page, C.L. Coleman, N. (1994) Metastar reduced chloride ion penetration. Aston University (data

sheet provided by ECCI).Page, C.L. and Havdahl, J. (1985) Electrochemical monitoring of corrosion of steel in microsilica

cement pastes. Materials and Structures, 18, No. 103, 41–47.Page, C.L. and Vennesland, O. (1983) Pore solution composition and chloride binding capacity of

silica fume cement pastes. Materials and Structures, 16, 19–25.Parker, D.G. (1986a) Alkali aggregate reactivity and condensed silica fume. Ref. No. SA 854/3C.

Elkem Materials Ltd.Parker, D.G. (1986b) Microsilica concrete. Part 2: In use. Concrete Society Current Practice Sheet

No. 110, Concrete, March.Paya, et al. (1986) Mechanical treatment of fly ashes, Part II. Cement and Concrete Research, 26,

No. 2, 225–235.Perry, C. and Gillott, J.E. (1985) The feasibility of using silica fume to control concrete expansion

due to alkali aggregate reaction. Durability of Building Materials, 3, 133–146.Pietersen, H.S. et al. (1993) Reactivity of flyash and slag in cement. Delft University.Pigeon and Regourd, M. (1983) Freezing and thawing durability of three cements with various

granulated blast furnace slag contents. Proc. Canment ACI Conference, ACI publication SP-79,V2, Montebello, Canada, 979–998.

Popovic, K, et al. (1984) Improvement of mortar and concrete durability by the use of condensedsilica fume. Durability of Building Materials, 2, 171–186.

Ramakrishnan, V. and Srinivasan, V. (1983) Performance characteristics of fibre reinforced condensedsilica fume concrete. American Concrete Institute, 11, 797–812.

Rasheeduzzafar et al. Proposal for a code of practice for durability of concrete in the Arabian Gulfenvironment.

Regourd, M. (1993) Condensed Silica Fume. In Aitcin, P.C. (ed.), Université de Sherbrooke,Canada, 20–24.

Cementitious additions 3/65

RMC Readymix Limited (1999) Unpublished data related to the Addashaw Farm project.Roy, D.M. (1987) Hydration of blended cements containing slag, fly ash or silica fume, Sir Frederick

Lea Memorial Lecture, 29 April–1 May 1987, Institute of Concrete Technology annual symposium.Ryle, R. (1994) Low permeability concrete for the Jubilee Line extension: An assessment of

proposed permeability tests, potential mixes and other factors. Confidential Technical Report257, RMC Readymix Limited.

Saad, M.N.A., de Andrade, W.P. and Paulon, V.A. (1982) Properties of mass concrete containing anactive pozzolan made from clay. Concrete International, July.

Samuelsson, P. (1982) The influence of silica fume on the risk of efflorescence on concrete surfaces.Norwegian Institute of Technology, Trondheim, Report BML 82.610, 235–244 (in Swedish).

Sandvik, M. (1981) Fasthetsutvikling for silicabetong ved ulike temperaturniva. FCB/SINTEF,Norwegian Institute of Technology, Trondheim, Report STF65 F81016 (in Norwegian).

Sandvik, M. (1983) Silicabetong: herdevarme, egenkapsutvikling. FCB/SINTEF, Norwegian Instituteof Technology, Trondheim, Report STF65 A83063 (in Norwegian).

Sellevold, E.J. and Radjy, F.F. (1983) Condensed silica fume (microsilica) in concrete: waterdemand and strength development. Publication SP-79. American Concrete Institute, 11, 677–694.

Sellevold, E.J. et al. (1982a) Silica fume-cement paste: hydration and pore structure. Report BML82.610. The Norwegian Institute of Technology, Trondheim, Norway, 19–50.

Sellevold, E.J. et al. (1982b) Silica Fume-cement pastes: hydration and pore structure. NorwegianInstitute of Technology, Trondheim, Report BML 82.610, 19–50.

Shirley, S.T. et al. Fire endurance of high strength concrete slabs. Concrete International.Shirvill, A.J. (1992) Mix design of metakaolin concrete, an interim report. Laboratory Report 23,

RMC Readymix Limited.Sibbick, R.G. and Nixon, P.J. (2000) Paper presented at 11th Int. Conf. on AAR, Quebec.Singh, B. and Osborne, G.J. (1994) Hydration and durability of OPC/Metakaolin blended concrete.

BRE Client Report CR291/94, BRE.Skurdal, S. (1982) Egenskapsutvikling for silicabetong ved forskjellige herdetemperaturer. Norwegian

Institute of Technology, Trondheim, Report BML 82.416 (in Norwegian).Smith, I.A. (1967) The design of fly ash concretes, ICE, Paper 6982, 769–790.Smolczyk, H.G. (1977) The use of blastfurnace slag cement in reinforced and prestressed concrete.

Sixth International Steel Making Day, Paris.Sorensen, E.V. (1982) Concrete with condensed silica fume. A preliminary study of strength and

permeability. Norwegian Institute of Technology, Trondheim, Report BML 82.610, 189-202 (inDanish).

Swamy, R.N. (ed.) (1986) Cement Replacement Materials, Surrey University Press. (Chapter 4written by Wainwright, P.J.)

Taylor, H.F.W. (1997) Cement Chemistry (2nd edn), Thomas Telford Publishing, London.Thomas, M.D.A (1990) A comparison of the properties of opc and fly ash concrete in 30 year old

mass concrete structures. In Durability of Building Materials and Components, E & FN Spon,London, 383–394.

Traetteberg, A. and Alstad, R. (1981) Volumstabilititet i blandingssementer med rajernslagg ogsilikastov. FCB/SINTEF, Norwegian Institute of Technology, Trondheim, Report STF65 A81034(in Norwegian).

Turrizani, R. (1964) Aspects of the chemistry of pozzolans. In The Chemistry of Cement, Chapter14, Vol. 2, Taylor, London.

Vennesland, O. (1981) Report 3: Corrosion properties. FCB/SINTEF, Norwegian Institute ofTechnology, Trondheim, Report STF65 A81031.

Vennesland, O. and Gjorv, O.E. (1983) Silica concrete – protection against corrosion of embeddedsteel. American Concrete Institute, publication SP-79, II, 719–729.

Virtanen, J. (1985) Mineral by-products and freeze–thaw resistance of concrete. Publikation nr.22:85. Dansk Betonforening, Copenhagen, 231–254.

3/66 Cementitious additions

Wainwright, P.J. and Aider H. (1995) The influence of cement source and slag additions on thebleeding of concrete. Cement and Concrete Research, Vol. 25. No. 7 1445–1456.

Wainwright, P.J. and Tolloczko, J.J.A. (1986) The early and later age properties of temperaturecycled opc concretes. 2nd Int. Conf on the use of fly ash silica fume, slag and natural pozzolansin concrete, CANMET/ACI April 1986, Madrid SP-91 Vol. 2, 1293–1321.

Walters, G.V. and Jones, T.R. (1991) Effect of metakaolin on alkali–silica reactions in concretemanufactured with reactive aggregates. Proc. 2nd Int. Conf. Durability of Concrete. Canada.

Wolsiefer, J. (1982) Ultra high strength field placeable concrete in the range 10 000 to 18 000 psi(69 to 124 MPa). American Concrete Institute, Annual Conference, Atlanta.

Wooley, G.R. and Cabrera J.G. (1991) Early age in-situ strength development of fly ash concretein thin shells, Int. Conf. on Blended cement, Sheffield.

Woolley, G.R. and Conlin, R.M. (1989) Pulverised fuel ash in construction of natural draughtcooling towers, Proc. Inst. Civil Engineers, Part 1 paper 9278, Vol. 86, 59–90.

Xu, H.Y. and Chen, M. (1986) AAR in Chinese engineering practices. Proc. 7th Int. Conf. on AAR.Ottawa.

Zhang, M.H. and Malhotra, V.M. (1995) Characteristics of thermally activated alumino-silicatepozzolanic material and its use in concrete. 2nd CANMET/ACI Int. Sym. On Advances in ConcreteTechnology.

PART 3

Admixtures

This Page Intentionally Left Blank

4

Admixtures for concrete,mortar and grout

John Dransfield

4.1 Introduction

4.1.1 Definition and description

Admixtures are chemicals, added to concrete, mortar or grout at the time of mixing, tomodify the properties, either in the wet state immediately after mixing or after the mixhas hardened. They can be a single chemical or a blend of several chemicals and may besupplied as powders but most are aqueous solutions because in this form they are easierto accurately dispense into, and then disperse through the concrete.

The active chemical is typically 35–40% in liquid admixtures but can be as high as100% (e.g. shrinkage-reducing admixtures) and as low as 2% (e.g. synthetic air-entrainingadmixtures). In most cases the added water from the admixture is not sufficient to requirea correction for water–cement ratio.

Admixtures are usually defined as being added at less than 5% on the cement in themix but the majority of admixtures are used at less than 2% and the typical range is 0.3–1.5%. This means the active chemicals are usually present at less than 0.5% on cement or0.02% on concrete weight.

The dosage may be expressed as litres or kg per 100 kg of cement, and cement normallyincludes any slag, pfa or other binders added at the mixer.

Admixtures are not the same as additives, which are chemicals preblended with thecement or a dry cementitous mix. Neither are they the same as additions, which are added

4/4 Admixtures for concrete, mortar and grout

at the mix. Type I additions are essentially inert, e.g. limestone powder or pigments. TypeII additions are pozzolanic or latent hydraulic binders such as pfa or silica fume.

4.1.2 Brief history of admixture use

Romans Retarders UrineAir entrainment BloodFibres Straw

Plasticizers 1932 Patent for sulphonated naphthalene formaldehydeplasticizers (but not available in commercialquantities)

193? Lignosulphonates used as plasticizers193? Hydroxycarboxcilic acid salts used as

plasticizers and retarders

Waterproofers 193? Fatty acids, stearates and oleates

Air entrainers 1941 Tallow and fatty acid soaps for frost resistance

Superplasticizers 1963 Sulphonated naphthalene formaldehydecommercially available

1963 Sulphonated melamine formaldehyde patentand available

1990–1999 Polycarboxylate ether development andintroduction

4.1.3 Admixture standards and types

Admixture standards in individual European countries were phased out during 2002 anda new European Standard EN 934 introduced, covering all the main types of admixture.This standard is currently divided into five parts. Part 2, covering concrete admixtures, isprobably the most important.

The previous British Standard was BS 5075 Parts 1 to 3 for Concrete and BS 4887 forMortars. Outside Europe, the American Standard ASTM C494 is widely specified.

See Figure 4.1 for admixture sales by type.

Admixture types covered by EN 934Normal plasticizing/water reducing (WRA) EN 934-2Superplasticizing/high range water reducing (HRWRA) EN 934-2Retarding & retarding plasticizing EN 934-2Accelerating-set and hardening types EN 934-2Air entraining EN 934-2Water retaining EN 934-2Water resisting (waterproofing) EN 934-2Retarded ready-to-use mortar admixtures EN 934-3Sprayed concrete EN 934-5Grout admixtures for prestressing EN 934-4

Admixtures for concrete, mortar and grout 4/5

Admixture types not covered by European standardsCorrosion inhibitingFoamed concrete & low density fill (CLSM)Polymer dispersionsPumping aidsSelf-compacting concretePrecast semi-dry concreteShrinkage-reducingUnderwater/anti-washoutWashwater

Figure 4.1 UK admixture sales by type (CAA 2000 statistics). These statistics show that normal plasticizersmade up the largest proportion of UK admixture sales. This is in contrast to much of Europe wheresuperplasticizers generally take a larger proportion of the market. This difference probably results from thepreference of UK buyers to specify 50 mm slump while most other countries specify slumps in excess of150 mm. Low slumps are generally difficult to place and fully compact and frequently lead to unrecordedaddition of water on the job site.

4.1.4 Admixture mechanism of action

Admixtures work by one or more of the following actions:

• Chemical interaction with the cement hydration process, typically causing an accelerationor retardation of the rate of reaction of one or more of the cement phases.

• Adsorption onto cement surfaces, typically causing better particle dispersion (plasticizingor superplasticizing action).

• Affecting the surface tension of the water, typically resulting in increased air entrainment.• Affecting the rheology of the water, usually resulting in an increased plastic viscosity

or mix cohesion.• Introducing special chemicals into the body of the hardened concrete that can affect

specific properties such as corrosion susceptibility of embedded steel or water repellence.

The benefits obtained from using the admixture are often in the hardened properties.However, with the possible exception of the last bullet point, these actions all affect theproperties of the wet concrete between the time of mixing and hardening in one or moreof the following ways:

• Affect water demand Plasticizing or water reducing• Change the stiffening rate Accelerating/retarding

All other8%

Plasticizers49%

Superplasticizers

12%

Mortars19%

Airentraining

5%

Waterresisting

7%Plasticizers

40%

All other13%

Waterresisting

2%Air

entraining2%

Mortars7%

Superplasticizers

36%

4/6 Admixtures for concrete, mortar and grout

• Change the air content Increase (or decrease) entrained air• Change the plastic viscosity Cohesion or resistance to bleed and segregation

of the mix

One of these effects will usually be the primary property, the property for which theadmixture is being used. However, the admixture can also affect one or more of the otherwet properties. These are called secondary effects and it is often these which are key toadmixture selection within an admixture type.

For example, a water-reducing admixture may be required to give low water:cementratio for durability but if the concrete mix lacks cohesion and is prone to bleed, a water-reducing admixture which also increases cohesion would be appropriate to prevent thebleed. The primary function is as a water reducer; the secondary function is to improvecohesion. Incorrect admixture selection could make the bleed worse.

Because of these secondary effects, an admixture which works well on one concreteplant with one source of cement and aggregates could be quite unsuitable on another planta few miles away where a different aggregate source is being used or a different bindercombination is employed.

4.1.5 Rheology and admixtures

The source of the materials used and the mix design determine the basic rheology of aconcrete mix. Only after this has been optimized should admixtures be used to modify therheology.

At the job site, concrete rheology is normally measured by the slump test and concretetechnologists may also use their experience to assess the cohesion. Is the mix too stickyor is it likely to segregate or bleed?

In the laboratory, as a measure of the admixture effect, we can use equipment to testfor the shear stress against rate of shear strain of a concrete mix. In concrete technologyterms the information results in a yield stress which is similar to a slump value and aplastic viscosity which puts a numerical value to cohesion.

From these rheology measurements we can gain an insight into the effect of individualmix constituents as shown in Figure 4.2. Adding water to a mix reduces yield (increasesthe slump) but also reduces plastic viscosity (cohesion) increasing the risk of bleed andsegregation. Dispersing admixtures (plasticizers and superplasticizers) will reduce the

Figure 4.2 Concrete rheology, effect of mix materials (P. Domone).

Less water

More pasteggbs

pfa

Cement-dispersing admixtures

Plastic viscosity

Morewater

Less pasteYield stress

Admixtures for concrete, mortar and grout 4/7

yield but may increase or decrease the plastic viscosity, depending on the secondaryproperties of the particular admixture chosen.

It is now possible to look at relative rheology that is required in concrete for variousapplications and where an admixture is typically used to achieve the requirements(Figure 4.3).

Figure 4.3 Concrete rheology by application (P. Domone).

Plastic viscosity

Self-compacting concrete

Pumping/underwaterconcrete

High-strength concreteNormalconcrete

Flowing concrete

Yield stress

High-strength concrete is often cohesive (high plastic viscosity) because of the highcement and low water content. Underwater and pump concrete also need to be cohesiveto resist segregation and admixtures are often used to achieve this. As the yield stressfalls, the plastic viscosity of normal concrete also falls unless the mix design is modifiedto increase cohesion and prevent segregation, as required by high-slump, flowing concrete.In self-compacting concrete, yield stress is almost zero but a range of plastic viscositiesis required to meet applications from a rapid flow, wet mix to a slow creeping flow withhigh plastic viscosity but still having a low yield stress.

Using this information the effects of individual admixtures or admixture combinationscan be selected to achieve the required changes in yield value and plastic viscosity for aparticular job function.

4.1.6 Cement and concrete chemistry in relation toadmixtures

Proportions of concrete mix constituents: By weight By volume

• Cement 15% 11%• Water required for cement hydration 3% 8%• Additional water to give workability 5% 11%• Aggregate 77% 70%

The most important factor in this typical mix design is the 11% by volume of water thatis added, only to give the mix workability. This water does not react with the cement butremains unbound and free in the concrete after hardening. This forms interconnectingwater-filled voids called capillaries that reduce strength, increase permeability to waterand provide paths for diffusion of chlorides, sulphates and other aggressive chemicals. Asthe water slowly evaporates from the capillaries, it causes drying, shrinkage and cracking.

4/8 Admixtures for concrete, mortar and grout

The empty capillaries provide a rapid path for carbonation.Water-reducing admixtures can significantly reduce the volume of this unbound water

leading to stronger and more durable concrete.Size of concrete mix constituents are typically as follows:

• Aggregate 20.0–4.0 mm• Sand 4.0–0.1 mm• Entrained air 0.3–0.05 mm• Cement 0.075–0.01mm• Capillaries < 0.001 mm• Gel pores <0.00003 mm• Admixture < 0.00001 mm

Entrained air voids are typically four times the size of a cement grain and are similarin size to the lower third of a typical sand grading. Air entrainment/stability and sandgrading/content are closely linked.

Capillaries are less than one tenth the diameter of a cement grain but account for 10%of the concrete volume – there must be a lot of them! Admixtures are less than a thousandththe diameter of a cement grain. They adsorb onto active sites on the grain.

Cement hydration starts as soon as water is added to the mix. Water absorbs into theouter part of the cement grain, dissolving calcium and hydroxide ions, which move outinto the surrounding water. This leaves the cement surface with a calcium-depletedhydrosilicate layer carrying a negative charge. Some of the positive calcium ions insolution then adsorb back onto this silicate surface to give a positive charge as measuredby the Zeta potential. Water continues to diffuse into the grain, releasing further calciumand hydroxide ions into solution and increasing the thickness of the hydrosilicate layer.

This is the predominant reaction between the cement and the water and mainly involvesthe tricalcium silicate (C3S) phase of the cement. However, other phases are also present,including tricalcium aluminate (C3A), which is very reactive although present at onlyabout 10% of the cement weight. Sulphate, in the form of gypsum, is added to the cementto help control the C3A reactivity by forming ettringite. Ettringite has a reactive surfacethat can attract and adsorb a disproportionate amount of some admixtures. For this reason,low C3A (sulphate-resisting or ASTM Type V cements) may require a slightly lowerdosage of admixture for equivalent effect.

Some admixtures can affect the rate of sulphate solubility, especially if the gypsumbecomes dehydrated to hemi-hydrate during cement grinding. This and some other changesin cement chemistry, which do not affect the cement properties on their own, can resultin abnormal setting in the presence of admixtures. This abnormal cement admixtureinteraction is very rare but embarrassing when it happens. It usually occurs soon after adelivery of very fresh cement.

When water is added to cement in a normal concrete mixer, the cement grains are notuniformly dispersed throughout the water but tend to form into small lumps or flocs.These flocs trap water within them and so the mix is less mobile and fluid than would bethe case if the cement were in the form of individual grains. High-shear grout mixers canbreak up these flocs producing a more fluid mix but there is a tendency for them to re-aggregate. Dispersing admixtures are used to both break up the flocs and to maintain thedispersion.

Admixtures for concrete, mortar and grout 4/9

4.2 Dispersing admixtures

Dispersing admixtures are the most important admixture type, accounting for over 60%of all admixtures sold in the UK. They include the types covered by the names Plasticizer,Superplasticizer, Water Reducer, High Range Water Reducer and they are covered byStandards EN 934-2 or by ASTM C494. Marketing ploys by admixture companies haveintroduced additional/alternative names including Mid-range, first, second and third generation,hyperplasticizer, etc. However, these names are generally not recognized by standards.

These dispersing admixtures all function by adsorption onto the cement surface in away that causes the cement particles to distribute more uniformly throughout the aqueousphase, reducing the yield value for a given water content and so increasing the fluidity ofthe mix. They may also have a small effect on the plastic viscosity or cohesion of the mix.They are used in one or more of the following ways:

• To increase the fluidity of the mix at a given water content. (Plasticizing orsuperplasticizing action)

• To reduce the water content of the mix at a given fluidity, improving strength anddurability. (Water reducing or High Range Water Reducing)

• To reduce the cement content by reducing water and maintaining water:cement ratioto give equivalent strength. (Cement reducing)

In optimizing a concrete mix for a given application it will often be effective to use acombination of these properties (for example, to use the admixture at a dosage which canreduce the water content but also allow the fluidity of the mix to increase).

Dispersing admixtures fall into to two groups which, for standards purposes (EN 934or ASTM C494), are divided by their water-reducing ability. These are the NormalPlasticizers/Water reducers and the Superplasticizers/High Range Water reducers.

4.2.1 Normal plasticizers

Definition: water reduction more than 5% but less than 12%.Other characteristics:

• Low dosage, typically 0.2–0.6% on cement (30–40% soln).• Often have significant secondary effects in the upper half of the dosage range, especially

retardation with some binder blends and/or at lower temperatures.• Higher dosages give little additional dispersion but very significant secondary effects.

Plasticizers are usually based on lignosulphonate, which is a natural polymer, derivedfrom wood processing in the paper pulp industry. The brown lignin in the wood pulp isremoved by a sulphite reaction and can then be processed to remove sugars and someother unwanted material before being used for admixtures. The greater the level of processing,the fewer the secondary effects.

Other dispersants may be blended with the lignin, e.g. gluconate, maltodextrin etc.These can boost certain properties but also tend to increase the potential for retardation.

The admixture may also be modified with air entrainers, air detrainers and smallamounts of other chemicals to boost or control secondary effects.

A more recent development is normal plasticizers based polycarboxylate ether. Thischemical was originally developed as a superplasticizer, but supplied as a dilute solution

4/10 Admixtures for concrete, mortar and grout

to decrease sensitivity; it can also be effective as a normal or multi range plasticizer andgives the advantage over lignosulphonates of having minimal secondary effects.

Normal plasticizers are extensively used by readymix companies to optimize their mixdesigns, especially at low to medium slump (up to about 130 mm). They account foralmost 50% of all admixture sales in the UK but a lower proportion in most othercountries where initial slump values tend to be significantly higher.

4.2.2 Superplasticizers

Definition: water reduction more than 12% but depending on dose and type can give over30%.

Other characteristics:

• Dosage, typically 0.6 to 2.0% on cement (30–40% soln).• Minimal retardation except at the top of the dosage range.• Synthetic chemicals derived from the chemical industry and designed with properties

optimized specifically for admixture applications. There are currently three main chemicaltypes: the sulphonated polymers of naphthalene or melamine formaldehyde condensatesand the polycarboxylate ethers. These will be discussed in more detail later in thissection.

These basic chemicals can be used alone or blended with each other or withlignosulphonate to give admixtures with a wide range of rheological properties. Theycan also be blended with other chemicals to increase retardation, workability retention,air entrainment and other properties.

Superplasticizers are a versatile group of admixtures with a wide range of potentialproperties and dosages, tailored to meet specific requirements. Dual-function productsthat are designed to give, say, water reduction and retardation are common, especially inwarmer climates. It is therefore essential to check with the data sheet and the manufacturerbefore moving from one type or brand of superplasticizer to another.

Superplasticizers find use in the pre-cast and readymix concrete industries as well asin site-mixed concrete. In pre-cast they are used mainly as water reducers to give highearly strength. On site-mixed they may be used:

• To give high workability where there is dense reinforcement• For low water contents to give high early or later age strength• For low w:c to give low-permeability durable concrete.

In readymix they are used for all the above applications and can also give extendedworkability to cope with long delivery and placing times, especially at elevated temperatures.

4.2.3 Chemical structure of dispersing admixtures

Lignin is a natural polymer found in wood and is of variable structure, depending on thetimber source and time of year when the wood is cut. In one paper pulping process, thebrown lignin is sulphonated to make it soluble prior to extraction as a by-product. Tomake this lignosulphonate suitable for use as a cement dispersant, it is then chemicallymodified by hydrolysis or fermentation to remove some of the carbohydrate material. Thekey part of the lignosulphonate can be regarded as sulphonated phenyl propane units of

Admixtures for concrete, mortar and grout 4/11

n

C

C

C

C

C

C

C

C

C

C

CH2

SO3Na

CH2

SO3 Na

C

C

C

C

C

C

C

C C

C

which a very simplified structure is shown in Figure 4.4. Potential ether linkages tofurther phenyl propane units to make up the polymer structure are indicated.

Figure 4.4 Simplified, sulphonated phenyl propane units showing some potential polymer crosslinkingpoints.

C

C

C

C

CH3O

C

CHO

CH

SO3

Na+

CH

OH

CH2OH

The polymeric lignosulphonate structure, represented in Figure 4.5, is a complexedand heavily crosslinked spherical microgel with a broad molecular weight distributionranging from about 2000 to 60 000. In solution the active sulphonate groups dissociate to– SO3

– and Na+ leaving the microgel with an overall negative charge which is key to itsdispersing properties.

Figure 4.5 Lignosulphonate microgel, less than one-thousandth the diameter of a cement particle.

• Active sulphonategroups

Sulphonated naphthalene formaldehyde condensate (SNF) also called polynaphthalenesulphonate (PNS) is derived from the chemical industry. Petroleum or coal tar naphthaleneis sulphonated using very concentrated sulphuric acid at high temperature and is thenpolymerized with formaldehyde and neutralized to the sodium (Na) or calcium (Ca) salt.The polymers are of relatively low molecular weight, the sulphonated naphthalene repeatingunit n shown in Figure 4.6 typically being in the range 2–10 to give molecular weights inthe order 500–2500. The higher molecular weights generally give better properties.

Figure 4.6 Sulphonated naphthalene formaldehyde condensate polymer.

The chain is linear and is free to rotate about the formaldehyde derived CH2 group sothe –SO3Na can be above or below the chain axis. In solution, like lignosulphonate, the

4/12 Admixtures for concrete, mortar and grout

CH2

z

CH CH2 CH

C O

O

Na

C O

O

CH2

CH2

O

CH3

n

–SO3Na dissociates to – SO3– and Na+. The negative charges on the – SO3

– are the key tothe admixture adsorbing onto the cement and also to its electrostatic dispersion mechanism.

Sulphonated melamine formaldehyde condensates (SMF) are very similar in structureexcept that a melamine ring replaces the naphthalene double ring and the molecularweight is higher (Figure 4.7). They are only available as the sodium salt.

Figure 4.7 Sulphonated melamine formaldehyde condensate polymer.

n

C

N

C

N

C

NH

CH2SO3Na

NH CH2NCH2 NHO

Polycarboxylate ethers (PCE) are also called PCs or comb polymers. They are themost recent development and unlike SNFs and SMFs, which are essentially a singlestructure, PCEs are a family of products with significantly different chemical structures.A typical PC structure is shown in Figure 4.8.

Figure 4.8 Basic structure of a polycarboxylate ether superplasticizer.

The backbone polymer is typically based on polymerization of acrylic acid but this canbe substituted or replaced with other monomer groups and can be used to modify thenumber of carboxylate groups along the polymer backbone. The carboxylate group isnormally neutralized as the sodium salt and takes on a negative charge in solution as theNa+ dissociates. This provides the attachment point for the admixture to adsorb onto thecement surface.

The co-polymer is a polyether shown here as polyethylene glycol (–CH2–CH2–O–)n.Other polyethers or combinations of polyether can be used and the molecular weightvaried, n typically ranging from 20 to 80 units. This together with the number of polyethergroups substituted along the chain and the length of the chain can give a significant rangeof properties. This allows the basic copolymer to be tailored for, say, high early strengthfor pre-cast or for workability retention for readymix even before other chemicals areblended in. The polyether is the part responsible for the dispersion of the cement particlesand works by a steric effect.

Admixtures for concrete, mortar and grout 4/13

4.2.4 Dispersing mechanism

When water is added to cement, the grains are not uniformly dispersed throughout thewater but tend to form into small lumps or flocs. These flocs trap water within themcausing the mix to be less mobile and fluid than would be the case if the cement were inthe form of individual grains. This is analogous to the situation where people are walkingboth ways along a narrow pavement. If groups of people walking in one direction holdhands, then it is hard for people walking the other way to pass, especially if they are alsoholding hands. If everyone stops holding hands and walking in groups, then it is easy forpeople going in opposite directions to move round each other and pass.

Admixtures adsorb onto the cement surfaces and break up the flocs, leaving individualcement grains, which can pass each other easily, making the mix more fluid (Figure 4.9).

Figure 4.9 Effect of dispersing admixtures in breaking up cement flocs.

Dispersing admixture

Absorbs on cementgrains

Cement uniformly dispersedincreasing fluidity

Cement flocs reducefluidity

There is no change in volume of the mix, just the release of water trapped within thefloc and the ability of individual cement grains to reorientate to make more efficient useof all the available volume without inter-particle contact. Not only has the fluidity increased,but with water surrounding the whole grain, hydration is more efficient.

If water is now removed from the system, the fluidity can be reduced back to what itwas before the admixture addition. However, the average inter-grain distance will alsoreduce, so less hydration will be necessary before the space between the cement grains isfilled with hydration product to give setting and strength development. This leads first toa higher early strength. With continued hydration, more inter-grain space eventuallybecomes filled with hydration products so late age strength is also improved. Finally,there will be less void space, not filled with any hydration product resulting in fewercapillaries and therefore better durability.

Admixture adsorption onto the cement surface is through the negative charges on theadmixture onto the positively charged calcium ions on the cement surface. The dispersionis then by one of two mechanisms, electrostatic repulsion or steric stabilization.

Electrostatic dispersion is the main mechanism in lignosulphonates and in SNF/SMFsuperplasticizers. These molecules all carry –SO3Na groups, which in water dissociateinto –SO3

– and Na+. The –SO3– remains attached to the admixture and carries a strong

negative charge. Part of this charge is used to attach the admixture to the cement but theremainder orientates out from the grain and repels the negative charges on admixtureadsorbed onto adjacent cement grains, causing them to move and stay apart. See Figure4.10.

As well as seeing the increased fluidity, the effect can be followed by looking at Zetapotential measurements. These measure the charge on the cement surface, which is positivebefore the admixture addition but then goes strongly negative as the admixture adsorbs.

4/14 Admixtures for concrete, mortar and grout

(See Figure 4.11). The magnitude of the change in the charge is much greater with thesuperplasticizers than with normal plasticizers, helping to explain the difference inperformance.

Figure 4.10 Electrostatic dispersion of cement grains by a SNF superplasticizing admixture. The realadmixture molecule is much smaller, relative to the cement grain, than is shown.

Figure 4.11 Typical effect of zeta potential (surface charge on cement grain) with addition of an SNFsuperplasticizer.

SNF dosage

0.4 0.8 1.2 1.6 2 2.4

–35

–30

–25

–20

–15

–10

–5

0

5

10

15

Zet

a po

tent

ial (

mV

)

Steric stabilization is the main mechanism in PCE superplasticizers. These moleculescarry –CO2Na groups which, in water, dissociate into –CO2

– and Na+. The –CO2– remains

attached to the admixture and carries a moderate negative charge that is used to attach theadmixture to the cement. The long polyether groups orientate away from the cementsurface but will resist becoming entangled with the polyether chains attached to anadjoining cement grain, thus keeping the two grains apart. Zeta potential measurementsshow that, unlike SNF, there is no significant negative charge on PCE superplasticizedcement (Figure 4.12).

Admixtures for concrete, mortar and grout 4/15

4.2.5 Normal plasticizer performance and applications

Admixture performance always depends on the concrete materials and mix design beingused as well as the dosage of the admixture. Normal plasticizers may be modified toenhance performance when slag or pfa binders are being used; they may also contain anair detrainer to control the air content of the mix. These and other modifications can affectcohesion, retardation, etc. and should be taken into account during admixture selection.

Typically, normal plasticizers will give 8–10% water reduction. Depending on retardationthis will translate to a 120% strength improvement at 24/36 hours and 112% at 28 days.It is difficult to improve on this water reduction but with a slightly higher dose, a higherworkability can be additionally obtained.

Most normal plasticizers will increase the air content by between 1% and 2%, whichwill give some additional cohesion to the mix and may reduce any tendency to bleedespecially if the plasticizer is being used to increase workability. The higher the aircontent of the mix without admixture, the greater will be the additional air that is entrainedby an admixture.

Most normal plasticizers give some retardation, 30–90 minutes over a control mix.However, this can become more significant with products that have been modified withmolasses, maltodextrin, hydroxylated polymers and other similar chemicals. The effectmay only become pronounced at higher dosages or lower temperatures and or wherehigher levels of a slag binder have been used. It can be a particular problem in slabs,where the low dosage of these admixtures contributes to non-uniform mixing and resultsin patches of concrete with different levels of retardation that are difficult to finish withpower-trowelling equipment and can lead to delamination in the hardened floor.

The main application of normal plasticizers is in readymix to optimize (reduce) thecement content by water reduction at constant w:c. They are also effective in providingincreased initial workability which, assisted by any increase in retardation, helps withworkability retention where long delivery times may occur.

Compared to a control mix without admixture:

• A normal plasticizer used for water reduction will give some improvement in otherproperties including permeability, chloride diffusion, shrinkage, etc.

Figure 4.12 Steric stabilization dispersion of cement grains by a PCE superplasticizing admixture. The realadmixture molecule is much smaller, relative to the cement grain, than is shown.

4/16 Admixtures for concrete, mortar and grout

• Used for cement reduction at constant/slightly reduced w:c the effects are more or lessneutral.

• Used only to increase workability, most properties including strength will be slightlyreduced.

4.2.6 Superplasticizer performance and applications

Superplasticizers can be supplied as pure products, SNF, SMF or PC, or they can beblended with each other or with lignosulphonate as well as with other modifying chemicals.In addition the concentration may vary to a much greater extent, 15–40%, than is usuallyfound in normal plasticizers which are more typically 30–42%. This means that the rangeof properties that can be obtained is much greater and the following notes are only ofgeneral guidance: there will be exceptions with some products.

Sulphonated melamine formaldehyde condensates (SMF) give 16–25%+ water reduction.They tend to reduce cohesion in the mix, increasing the tendency to bleed and segregatebut can be very effective in improving mixes which, because of the materials/mix design,are already over-cohesive and sticky. SMF gives little or no retardation, which makesthem very effective at low temperatures or where early strength is most critical. However,at higher temperatures, they lose workability relatively quickly and are often added at thesite rather than during batching at a readymix plant. SMF generally give a good finish andare colourless, giving no staining in white concrete. They are therefore often used whereappearance is important.

Sulphonated naphthalene formaldehyde condensates (SNF) typically give 16–25%+water reduction. They tend to increase the entrapment of larger, unstable air bubbles. Thiscan improve cohesion but may lead to more surface defects. Retardation is more thanwith SMF but will still not normally exceed 90 minutes over a control mix, even at thehighest dosages. Workability retention is quite good but can be significantly improved ifmodified with other plasticizers/retarders. SNF is a very cost-effective and forgivingsuperplasticizer and has found widespread use in all parts of the industry especially inreadymix and site-mix applications.

Polycarboxylate ether superplasticizers (PCE) typically give 20–35%+ water reduction.They are relatively expensive per litre but are very powerful so a lower dose (or moredilute solution) is normally used. As previously noted, the basic molecule can be manipulatedto give targeted properties. This has currently led to products targeted at pre-cast, havingvery good water reduction, little retardation but only moderate workability retention.Other formulations target readymix with enhanced workability retention but more normallevels of water reduction and retardation. The PCEs are the most recent superplasticizerdevelopment and are still establishing their position in the market, with developmentcontinuing to optimize their performance. They have already gained a dominant positionin the pre-cast market and for self-compacting concrete but can be more expensive andless forgiving of mix design and material consistency than SNF for general readymix andsite use, but even here the PCE market share is growing.

The main benefits of superplasticizers can be summarized as follows:

• Increased fluidity for: Flowing, self-levelling, self-compacting concretePenetration and compaction round dense reinforcement

Admixtures for concrete, mortar and grout 4/17

• Reduced w:c for: Very high early strength, >200% at 24 hours or earlierVery high later age strengths, >100 MPa.Reduced shrinkage, especially if combined with reducedcement content.Improved durability by removing water to reduce permeabilityand diffusion

Superplasticizers are not as cost effective as cement reducers.

4.3 Retarding and retarding plasticizing/superplasticizing admixtures

Definition: extend the time for the mix to change from the plastic to the hardened state byat least 90 minutes but not more than 360 minutes at the compliance dosage.

If also plasticizing/water reducing then should also give more than 5% but normallyless than 12% water reduction. If also superplasticizing/high range water reducing thenshould also give more than 12% water reduction but, depending on dose and type, cangive significantly more.

Dosage is typically 0.2–0.6% on cement if only supplied as a retarder. If it is amultifunction admixture with dispersing properties, the dosage will be in the same rangeas for plasticizing/superplasticizing admixtures.

Most retarding admixtures are sold as multifunction retarding plasticizing/superplasticizing products. They are based on the same chemicals as the dispersingadmixtures but have molasses, sugar or gluconate added to provide the retardation. Pureretarders are usually based on gluconate and or sucrose although these also plasticize.Phosphates are occasionally used and retard with little plasticizing action.

Significant overdosing of most retarding admixtures quickly lead to large extensionsof set time running into days and in extreme cases can result in the cement never properlyhydrating and gaining strength. Phosphates are more tolerant to overdosing but are lesseffective in most applications and are rarely used.

Temperature has a significant effect on retardation and this needs to be taken intoaccount when selecting the dosage or running trial mixes. Even a 5° temperature changecan have a large effect. Retardation also increases if pfa or ggbs form part of the cementcontent and should also be taken into account, especially if the temperature is likely to below.

Retarding admixtures find use in readymix and on-site concrete to help provideworkability retention and/or to prevent cold joints on large or delayed pours.

4.3.1 Mechanism of retardation

The mechanism of setting of cement is not fully understood and this impacts on ourunderstanding of the way retarding admixtures work. There are probably two main processesinvolved:

• A blocking mechanism where the admixture adsorbs strongly on the cement surface,slowing the formation of silicate hydrates.

4/18 Admixtures for concrete, mortar and grout

• Chelation of calcium ions in solution, preventing the precipitation of calcium hydroxide(portlandite).

One or both of these processes may be involved, depending on the admixture typeselected. The latter is probably the more important for most retarder types and preventssetting but not workability loss. This is important to appreciate if the required performanceis to be achieved from a retarding admixture.

4.3.2 Workability retention

Cement hydrates by progressive absorption of water into the grain followed by the formationof calcium silicate hydrates (CSH). The total water required for hydration is typically inthe order of 70–90 litres per cubic metre of concrete. While some of this water will absorbduring the initial mixing, further absorption occurs during the plastic stage. Each 20 litresof water lost into the cement grain will roughly half the workability. Combine this withthe increased interaction between the hydration products being formed and it is apparentwhy there is progressive workability loss.

Retarding admixtures do not slow the rate of water absorption and do not prevent allthe hydration reactions so they have little effect on workability loss.

The best way to have the required workability at a given time after mixing is to use aplasticizer/superplasticizer to increase the initial workability to a point where the workabilityloss will give the required slump at the required time (Figure 4.13).

Figure 4.13 Initial workability of concrete and typical workability loss with time.

Control

Retarded plasticized

Retarded superplasticized

Plasticized

0 30 60 90 120 150 180 210 240 270Time after mixing (min)

250

200

150

100

50

0

Slu

mp

(mm

)

If the high workability is required close to or after the time of initial set of theconcrete, then a retarder will be required in addition to the plasticizer in order to slowdown the hydration and setting reactions of the cement. However, as many plasticizersalso slightly retard as a secondary effect, there may be no requirement for the retardingadmixture. In Figure 4.13 if a slump of 100 mm were required at 90 minutes then anormal plasticizer would probably meet the requirements.

Admixtures for concrete, mortar and grout 4/19

4.3.3 Set retardation

When water is first added to cement there is a rapid initial hydration reaction, after whichthere is little formation of further hydrates for typically 2–3 hours. The exact time dependsmainly on the cement type and the temperature. This is called the dormant period whenthe concrete is plastic and can be placed. At the end of the dormant period, the hydrationrate increases and a lot of calcium silicate hydrate and calcium hydroxide is formedrelatively quickly. This corresponds to the setting time of the concrete.

Retarding admixtures delay the end of the dormant period and the start of setting andhardening. This is useful when used with plasticizers to give workability retention. Usedon their own, retarders allow later vibration of the concrete to prevent the formation ofcold joints between layers of concrete placed with a significant delay between them.

4.3.4 Retarder performance and applications

The use of retarding admixtures to assist plasticizers/superplasticizing admixtures ingiving workability retention over extended periods beyond the normal setting time of themix is well covered above. Workability retention is required in order to place and fullycompact the concrete.

Cold joints are formed when fresh concrete is placed onto a previous pour that hasalready undergone an initial set, preventing the two pours being vibrated into a singlemonolith. It typically occurs on large or complicated pours when, after placing onesection of concrete, adjacent areas of concrete are placed over an extended period beforefurther concrete is placed onto or adjacent to the first pour. Over this time period, the firstconcrete pour has not only lost workability but has started to set so that it is no longeraffected by the action of a vibrator. The new batch of concrete fails to fully bond to theearlier pour and a line of weakness with respect to both strength and permeability nowexists in the structure. The first batch of concrete does not need workability for the twopours to be vibrated but it must not have gone beyond its initial set. A retarder can be usedto achieve this, and in extreme cases can delay the initial set for over 24 hours for verydifficult or large pours.

Cold joints often occur as a result of delays and breakdowns on-site and this possibilityshould be taken into account when planning any large job. Readymix supply can beinterrupted by a traffic accident or by a plant breakdown, on-site a pump blockage orbreakdown can also cause a serious delay in the placing of concrete. If this results in acold joint in a bridge deck or a slip form, the resulting problems can be extremelyembarrassing.

Mass concrete pours can also cause problems with cold joints. Unretarded concrete,placed in the early stages of the pour, will start to set and exotherm after a few hours. Theheat liberated will accelerate the set of the concrete above and can eventually affect theactive surface, resulting in an unexpectedly quick set and formation of a cold joint. Theuse of a retarder in the early pours with a progressively reduced dosage as the pourprogresses can solve this problem.

Retarders do not reduce the exotherm or the peak temperature of mass concrete. Theyonly delay the onset of the exotherm. The only effective way to control exotherm is by theuse of less reactive binders such as pfa or ggbs. Chilled mixing water or cooling theaggregate are also effective ways of reducing the peak temperature.

4/20 Admixtures for concrete, mortar and grout

Accidental overdosing of a retarding admixture can significantly delay the setting andstrength development. As soon as the error is spotted, the following actions should betaken:

• Protect the surface from evaporation by effective curing• Revibrate the concrete up to the time of initial set to prevent settlement cracking• Contact the admixture manufacturer for advice on the likely retardation time at the

actual dosage used.

Unless an extreme overdose was used, most concrete will set and gain strength normallywithin 3–5 days. Beyond 5 days, the potential for the concrete not to gain full strengthincreases significantly.

4.4 Accelerating admixtures

Accelerating admixtures can be divided into groups based on their performance andapplication:

• Set accelerating admixtures, reduce the time for the mix to change from the plastic tothe hardened state. They can be subdivided into two groups:– Sprayed concrete accelerators, which give very rapid set acceleration (less than 10

minutes) and will be covered in section 4.11.– Concrete set accelerators which, according to EN 934-2, reduce the time for the

mix to change from the plastic to the hardened state by at least 30 minutes at 20°Cand at least 40% at 5°C.

• Hardening accelerators, which increase the strength at 24 hours by at least 120% at20°C and at 5°C by at least 130% at 48 hours.

Other characteristics are:

• Dosage, typically 0.5–2.0% on cement.• Most accelerators are inorganic chemicals and in most cases are either set or hardening

accelerators but not both, although blends are available which provide a degree of bothproperties.

Calcium chloride is the most effective accelerator and gives both set and hardeningcharacteristics. Unfortunately, it also reduces passivation of any embedded steel, leadingto potential corrosion problems. For this reason, it should not be used in concrete whereany steel will be embedded but may be used in plain unreinforced concrete.

Chloride-free accelerators are typically based on salts of nitrate, nitrite, formate andthiocyanate. Hardening accelerators are often based on high range water reducers, sometimesblended with one of these salts.

Accelerating admixtures have a relatively limited effect and are usually only cost-effective in specific cases where very early strength is needed for, say, access reasons.They find most use at low temperatures where concrete strength gain may be very slowso that the relative benefit of the admixture becomes more apparent.

Admixtures for concrete, mortar and grout 4/21

4.4.1 Mechanism of acceleration

Like retarders, the mechanisms are not well understood. Set accelerators appear to acceleratethe formation of ettringite. Inorganic hardening accelerators increase the rate of dissolutionof the tricalcium silicate, leading to an increase in calcium silicate hydrate (CSH) at earlyages. Hardening accelerators based on high range water reducers reduce the distancebetween cement grains so that, for a given amount of CSH hydration product, there ismore interaction between the cement grains and hence more strength.

4.4.2 Accelerator performance and applications

Set accelerators have relatively limited use, mainly to produce an early set on floors thatare to be finished with power tools. They can, for instance, be used only in the later mixesso that these are accelerated and reach initial set at the same time as earlier pours,allowing the whole floor to be finished at the same time.

Hardening accelerators find use where early stripping of shuttering or very earlyaccess to pavements is required. They are often used in combination with a high rangewater reducer, especially in cold conditions.

At normal ambient temperatures, hardening accelerating admixtures are rarely cost-effective and a high range water reducer is usually a more appropriate choice. The relativeeffects are shown in Table 4.1.

Table 4.1 Comparison between accelerating admixtures and a high range water reducer

Temperature 5°C 20°CAccelerator Accelerator High range water

reducer

Final set, hours 5.5 2.25 4.756 hours, strength N/mm2 0.0 7.0 2.58 hours 1.8 12.5 8.512 hours 3.5 20.0 23.524 hours 12.5 27.5 35.028 days, strength N/mm2 48.5 50.5 63.0

In summary, a hardening accelerator may be appropriate for strength gain up to 24hours at low temperature and up to 12 hours at ambient temperatures. Beyond thesetimes, a high range water reducer alone will usually be more cost-effective.

4.5 Air-entraining admixtures

These are defined in EN 934-2 as admixtures that allow a controlled quantity of small,uniformly distributed air bubbles to be incorporated during mixing and which remainafter hardening. Air entrainment is usually specified by the percentage of air in the mixbut may also be specified by air void characteristics of spacing factor or specific surfaceof the air in the hardened concrete.

4/22 Admixtures for concrete, mortar and grout

Dosage is typically 0.2–0.4% on cement weight unless they are supplied as amultifunctional product, usually with a dispersing admixture. Air-entraining admixturesare very powerful and may only contain 1–5% solids. The low dose and powerful actionrequire good, accurate and well-calibrated dosing equipment if a consistent level of air isto be achieved.

Strength is reduced by entrained air, typically 5–6% reduction for each 1% of additionalair. For this reason, air should normally be limited to the lowest level necessary to achievethe required properties. However, in low cement content, harsh mixes, air may improveworkability allowing some water reduction that can offset part of the strength loss.

Air entrainment is used to produce a number of effects in both the plastic and thehardened concrete. These properties will be considered in more detail later in the sectionbut include:

• Resistance to freeze–thaw action in the hardened concrete• Increased cohesion, reducing the tendency to bleed and segregation in the plastic

concrete• Compaction of low workability mixes including semi-dry concrete• Stability of extruded concrete• Cohesion and handling properties in bedding mortars

Air entrainment is very sensitive to the concrete mix constituents, mixing, temperature,transport, pumping etc. and achieving consistent air content from batch to batch can bedifficult if these parameters are not well controlled.

Air-entraining admixtures are surfactants that change the surface tension of the water.They have a charged hydrophilic end that prefers to be in water and a hydrophobic tail,which traps air bubbles as a way of getting out of the water. Traditionally, they were basedon fatty acid salts or vinsol resin but these have largely been replaced by syntheticsurfactants or blends of surfactants to give improved stability and void characteristics tothe entrained air.

4.5.1 Factors affecting air entrainment

The sand content and grading undoubtedly has the greatest influence on air entrainment.It is almost impossible to entrain a significant amount of air into a neat cement paste(except by the use of a pre-foam system). As the sand content increases, so does theamount of air under otherwise equal conditions, with an optimum level being reached atabout 1:3 cement to sand. Entrained air is very similar in size to the bottom third of mostsand gradings (see relative size table in section 4.1.6) and is probably stabilized byslotting into appropriate sized holes in the grading curve.

Sand grading also has a significant effect as can be seen in a control mix with noadmixture. For a given mix design, different sand sources with different gradings canchange the air content from about 0.6% to over 3%. The more a sand tends to entrain air,the higher the air content that will be entrained by an admixture at a given dose (note thatthis also applies to plasticizers that may have air entrainment as a secondary effect).

Cement fineness and dust in aggregates also affect stability, probably by clogging upgaps between sand grains, reducing suitable voids for bubble formation and by absorptionof free water from the bubble surfaces.

Admixtures for concrete, mortar and grout 4/23

In relation to production of consistent air-entrained concrete, changes in cement contentand fineness should be avoided and sand grading should also be consistent. Dust in anyof the aggregates should be avoided. Because air is similar in size to the finer sand, it actsin a similar way, increasing cohesion. In already cohesive mixes (high cement content orfine sand) it may be necessary to reduce the sand content of the mix to reduce cohesionbut if the cement:sand ratio falls to about 1:2 problems with consistency and stability ofair content may be encountered.

The use of pfa in air-entrained concrete can give significant problems in achieving therequired air content and batch-to-batch consistency, often resulting in low air contentsand/or much higher than expected admixture dose. This is due to the presence of partiallyburnt coal, which has an active surface that adsorbs the surfactants used as air-entrainingadmixtures. The ‘Loss on Ignition’ figure for the pfa may not be a guide to potentialproblems as this test also picks up unburned coal, which does not adsorb the surfactant.Some types of air-entraining admixture are more tolerant of pfa than others but none areimmune to the problem, so great care and regular checking is essential, especially whentaking delivery of a fresh batch of pfa.

Entrained air is reasonably tolerant of continued agitation but can be lost if subjectedto delivery in a non-agitated vehicle such as a tipping lorry. Air can also be lost duringpumping. It is therefore best to check the air content at the final point of discharge intothe shutter and adjust the admixture to achieve the required amount of air at that point.Finally, over-vibration can also reduce the air content and may additionally result in afoam of air on the top surface. It should be noted that the higher the workability of theconcrete, the greater is likely to be the loss of air from any of these factors.

4.5.2 Freeze–thaw resistance

Frost damage to concrete results from water within the capillary system freezing andexpanding to cause tensile forces which crack the concrete or cause surface scaling. Itshould be remembered that the capillaries can amount to about 10% of the concretevolume and are the result of the excess water added to the concrete to give workability,see section 4.1.6. For frost damage to occur, the concrete must be almost saturated(capillaries full of water).

In air-entrained concrete, air voids are formed that intersect the capillaries at regularintervals, preferably at no more than 0.4 mm spacing. This equates to a void spacingfactor of 0.2 mm (the maximum distance from any point in the capillary system to thesurface of the nearest air bubble). To achieve this at a 5% air content, most of the airbubbles must be between 0.3 and 0.1 mm diameter. This will give a specific surface to thebubble of between 20 and 40 mm–1. If one air bubble of 10 mm diameter gave 5% air ina concrete mix, it would take 300 000 bubbles of 0.1 mm diameter to give a specificsurface of 40 mm–1 at the same air content. These air voids are still about a hundred timeslarger than the capillaries.

When the concrete becomes wet, water will be drawn in by capillary suction but anywater that enters an air void will prefer to be sucked back into the capillary system. Thismeans that in most situations the air voids remain dry. The exception is when the waterenters the concrete under pressure. If the water in the capillaries starts to freeze, thepressure in the capillary system rises but before it reaches a level which would crack the

4/24 Admixtures for concrete, mortar and grout

concrete, unfrozen water is forced into the air voids and the pressure is released. Theeffectiveness of this system is demonstrated in Figure 4.14 where even the minimum aircontent of 3.5% is enough to dramatically increase the number of freeze–thaw cyclesbefore serious damage occurs to the concrete. This is compared to a high-strength-waterreduced concrete, which shows significant distress after only about 60 freeze–thaw cycles.

ControlHRWRAEA

0 30 60 90 120 150 800 1000Freeze–thaw cycles

120

100

80

60

40

20

0

Dyn

amic

mod

ulus

(%

)

Figure 4.14 ASTM C666 freeze–thaw test on concrete prisms shows the effectiveness of air entrainment.

4.5.3 Air entrainment to reduce bleed

Bleed is usually a function of sand grading and is difficult to overcome. Reducing watercontent/workability or increasing sand are relatively ineffective and an alternative sandsource may not be cost-effective but a small amount of entrained air can make a significantdifference.

Bleed can result in sand runs or water lenses under coarse aggregate that show assurface defects. Paste or grout may wash out if there are any holes in shutters. On thesurface a weak layer of high w:c paste and latence may form. Bleed is actually a functionof sand settlement and may lead to plastic settlement cracking, which forms from the toplayer of reinforcement up to the surface and is a serious durability risk.

One to two per cent of additional air can very significantly reduce most bleed problemswithout having an over-serious effect on strength. Indeed, some plasticizers will put inthis level of air as a secondary effect and may be sufficient to deal with the problem. Thebenefit from the air probably comes from its ability to form air voids that match the gapsin the sand grading, filling the spaces and reducing the opportunity for settlement.

4.5.4 Compaction and cohesion of low-workability mixes

Mixes with very low workability (zero slump), and/or which have a low cement contentand are harsh can be very difficult to place and compact. In these mixes, the presence ofsome entrained air can greatly enhance the compaction under vibration. The air helps tolubricate contact points where the aggregates touch and would otherwise prevent normalflow and compaction. In some cases air may be formed at these points under the actionof the vibrator and be lost again when vibration stops. The admixture helps the mix tocompact under vibration, closing voids and forming a good surface but the mix stiffensagain very quickly when vibration stops, preventing further slumping. This helps to

W/C Air N/mm2

Control 0.50 1.3 58HRWR 0.40 1.0 74AEA 0.47 3.5 53

Admixtures for concrete, mortar and grout 4/25

preserve the shape of the unit, especially where a free vertical surface is formed inextruded or instantly demoulded units.

Typical applications include: horizontal slip-forming, extruded slabs, blinding, screeds,semi-dry bricks and blocks.

4.5.5 Bedding mortars and renders

Air is extensively used in bedding mortars and renders as a ‘plasticizer’ to reduce watercontent and give a softer feel during trowelling. It also gives some protection against frostaction, even at an early age and helps to control strength in richer mixes. The air contentis much higher than that used in concrete, being typically in the range 14–20%. Admixturesfor this application are covered by EN 934-3.

4.6 Water resisting (waterproofing)

These are defined in EN 934-2 as admixtures that reduce the capillary absorption of waterinto the hardened concrete. There are, in addition, ‘permeability-reducing’ admixturesthat reduce the passage of water through the concrete under a pressure head. Most productsfunction in one or more of the following ways:

• Reducing the size, number and continuity of the capillary pore structure• Blocking the capillary pore structure• Lining the capillaries with a hydrophobic material to prevent water being drawn in by

absorption/capillary suction.

All these ‘waterproofing’ admixtures reduce surface absorption and water permeability ofthe concrete by acting on the capillary structure of the cement paste. They will notsignificantly reduce water penetrating through cracks or through poorly compacted concretewhich are two of the more common reasons for water leakage through concrete.

Dosage varies widely, depending on the type and the required level of performance.Two per cent is common for the hydrophobic types, but may be 5% or more for the poreblockers. Dosage is often given as a weight or volume per cubic metre rather than oncement weight as their action is largely independent of cement content.

Lowering the free water content of the mix reduces the size and continuity of capillarypores and this can be achieved with water-reducing or high range water-reducing materials.Reducing capillary size alone does not necessarily reduce surface absorption and mayeven increase it as capillary suction increases with smaller diameter pores. However, theoverall pressure permeability of the concrete will be improved due to greater capillarydiscontinuity.

Pore blocking can be achieved by the addition of very fine unreactive or reactiveadditions such as silica fume or by the use of insoluble organic polymers such as bitumenintroduced as an emulsion. The hydrophobic admixtures are usually derivatives of long-chain fatty acid of which stearate and oleate are most commonly used. These may besupplied in liquid or powder forms. Some admixtures are combinations of two or more ofthese systems.

4/26 Admixtures for concrete, mortar and grout

4.6.1 Mechanism

Water will be absorbed, even into dense concrete, through the capillary pores. However,if the added mix water can be reduced to a w:c of below about 0.45 (or 160 litres per cubicmetre of added water) the capillary system loses much of its continuity and pressurepermeability is significantly reduced. The best way of achieving this reduced level ofwater is to use a water-reducing admixture that can also be used to ensure sufficientworkability for full compaction and reduce shrinkage cracking.

The pore-blocking admixtures are based on very fine reactive or unreactive fillers orinsoluble polymer emulsions, which have particle sizes of around 0.1 microns and aresmall enough to get into the capillaries during the early stages of hydration and physicallyblock them.

The hydrophobic admixtures are usually designed to be soluble as an admixture butreact with the calcium of the fresh cement to form an insoluble material which adsorbsonto the surfaces of the capillaries. Once the capillary dries out, the hydrophobic layerprevents water re-entering the capillary by suction but resistance is limited and dependsupon the head of water involved, the quality of the concrete and the effectiveness of theadmixture.

4.6.2 Admixture selection

Hydrophobic ‘water-resisting admixtures are generally effective:

• against rain• against surface water• against low-pressure heads in structures• water ingress in tidal and splash zone• reducing chloride diffusion by preventing build-up of absorbed chloride at the surface.

They are not effective against water with a continuous pressure head.Pore blockers are best if used in combination with a high range water reducer. They are

effective in reducing water penetration under a pressure head of several atmospheres andpermeability coefficient can be improved from typically 10–10 to better than 10–13 m/s.

Most waterproofing admixtures tend to slightly reduce the strength of concrete, eitherdue to some increased air entrainment or to reduced cement hydration. Integral waterreduction is often enough to offset this strength loss but some admixtures, especially theliquid hydrophobic types, may need the separate addition of a water reducer.

4.7 Corrosion-inhibiting admixtures

Corrosion-inhibiting admixtures are not currently covered by EN 934 but are effectiveafter the concrete has hardened and give a long-term increase in the passivation state ofsteel reinforcement and other embedded steel in concrete structures. The three mostcommon generic types of corrosion inhibiting admixture are:

• Calcium nitrite (normally contains a residual amount of calcium nitrate)

Admixtures for concrete, mortar and grout 4/27

• Amino alcohols• Amino alcohols blended with inorganic inhibitors

The dosage of corrosion inhibitors is usually dependent upon the client’s expected serviceablelife of the structure and on a range of factors that affect the durability of concrete. Theseinclude cement type, water-to-cement ratio, cover concrete to the steel, ambient temperatureand the expected level of exposure to chlorides. The typical dosage range for a 30%solution of calcium nitrite is 10–30 litres/m3 but is more usually used between 10 and 20litres/m3. The dosage of amino alcohol-based corrosion inhibitors is usually between 3and 4 volume per cent by weight of cement. Both types can be used with other admixturetypes and their use with a high range water-reducing admixture is usually recommendedin order to ensure the quality and durability of the base concrete.

4.7.1 Mechanism

The most common cause of reinforcement corrosion is pitting corrosion due to the ingressof chloride ions through the covering concrete and subsequent diffusion down to theembedded steel. When the amount of chloride at the steel surface reaches a critical level,the passivation of the steel breaks down and corrosion starts. Carbonation of the concreteleads to a lowering of the alkalinity around the steel and when the pH drops below acritical level this causes a loss of passivation that results in general reinforcement corrosion.

Corrosion inhibitors increase the ‘passivation state’ or ‘corrosion threshold’ of thesteel so that more chloride must be present at the steel surface before corrosion can start.Although corrosion inhibitors can raise the corrosion threshold, they are not an alternativeto using impermeable, durable concrete and are not cost-effective unless used in a high-quality mix.

The mechanism by which corrosion inhibitors operate is dependent on their chemicalnature. Calcium nitrite-based corrosion inhibitors convert the passive layer on the steelsurface into a more stable and less reactive state. When the chloride ions reach the layer,no reaction occurs – the steel is in a passive state. It is the anodic corrosion sites on thesteel that are protected against the chloride attack and for this reason nitrites are calledanodic inhibitors. It is important that sufficient nitrite is present to counter the chlorideions. The dosage of the admixture is therefore based on the predicted level of chloride atthe steel over the design life of the structure.

Amino alcohol-based corrosion inhibitors coat the metal surface with a monomolecularlayer that keeps the chloride ions away from the embedded steel. They also inhibit thereaction of oxygen and water at the cathodic sites on the steel, which is an essential partof the corrosion process. As a result amino alcohols can be regarded as both anodic andcathodic inhibitors.

4.7.2 Use of corrosion inhibitors

Corrosion inhibitors can significantly reduce maintenance costs of reinforced concretestructures throughout a typical service life of 40–60 years. Structures especially at riskare those exposed to a maritime environment or other situations where chloride penetration

4/28 Admixtures for concrete, mortar and grout

of the concrete is likely. Such structures include jetties, wharves, mooring dolphins andsea walls. Highway structures can be affected by the application of de-icing salts duringwinter months, as can multi-storey car parks where salt-laden water drips off cars andevaporates on the floor slab.

The concrete should be designed with a low diffusion coefficient and sufficient coverto slow the rate of chloride reaching the steel, the corrosion inhibitor can then extend thetime to onset of corrosion as indicated by the following example:

Concrete spec.PC 400 kg/m3 w:c 0.40Reinforcement at 40 mmChloride diffusion coeff. 1 × 10–12 m2/s

No corrosion inhibitorChloride corrosion threshold 0.4% on cementThreshold reached after 20 years

Calcium nitrite added at 10 litres/m3 (30% soln)Chloride corrosion threshold raised to 0.8% on cementThreshold reached after 50 years

In this example, the inhibitor has not reduced the amount of chloride that reaches the steelbut has increased the amount of chloride that must be present at the steel surface beforecorrosion is likely to start.

4.8 Shrinkage-reducing admixtures

Shrinkage-reducing admixtures can significantly reduce both the early and long-termdrying shrinkage of hardened concrete. This is achieved by treating the ‘cause’ of dryingshrinkage within the capillaries and pores of the cement paste as water is lost. This typeof admixture should not be confused with shrinkage-compensating materials which arenormally added at above 5% on cement and function by creating an expansive reactionwithin the cement paste to treat the ‘effects’ of drying shrinkage.

Shrinkage-reducing admixtures are mainly based on glycol ether derivatives. Theseorganic liquids are totally different from most other admixtures, which are water-basedsolutions. Shrinkage reducing admixtures are normally 100% active liquids and are water-soluble. They have a characteristic odour and a specific gravity of less than 1.00. Thedosage is largely independent of the cement content of the concrete and is typically in therange 5–7 litres/m3.

4.8.1 Mechanism

When excess water begins to evaporate from the concrete’s surface after placing, compacting,finishing and curing, an air/water interface or ‘meniscus’ is set up within the capillariesof the cement paste. Because water has a very high surface tension, this causes a stress tobe exerted on the internal walls of the capillaries where the meniscus has formed. Thisstress is in the form of an inward-pulling force that tends to close up the capillary. Thus

Admixtures for concrete, mortar and grout 4/29

the volume of the capillary is reduced, leading to shrinkage of the cement paste aroundthe aggregates and an overall reduction in volume of the concrete.

The shrinkage-reducing admixtures operate by interfering with the surface chemistryof the air/water interface within the capillary, reducing surface tension effects andconsequently reducing the shrinkage as water evaporates from within the concrete. Theymay also change the microstructure of the hydrated cement in a way that increases themechanical stability of the capillaries.

4.8.2 Use of shrinkage-reducing admixtures

Shrinkage-reducing admixtures can be used in situations where shrinkage cracking couldlead to durability problems or where large numbers of shrinkage joints are undesirable foreconomic or technical reasons. In floor slabs, the joint spacing can be increased due to thereduced movement of the concrete during drying. The risk of the slab curling at jointsand/or edges is also significantly reduced. Where new concrete is used to strengthen orrepair existing structures, shrinkage-reducing admixtures can reduce the risk of crackingin what can be a highly restrained environment.

4.9 Anti-washout/underwater admixtures

Underwater concrete anti-washout admixtures are water-soluble organic polymers whichincrease the cohesion of the concrete in a way that significantly reduces the washing outof the finer particles, i.e. cementitious material and sand from fresh concrete when it isplaced under water. They are often used in conjunction with superplasticizers to produceflowing self-levelling concrete to aid placing and compaction under water.

Anti-washout admixtures were developed to provide a higher integrity of concreteplaced under water and to reduce the impact that the washed-out material can have on themarine environment. They are suitable for use in deep underwater placement, in inter-tidal zones and in the splash zone and other situations where water movement may resultin the cement and other fine material being washed out.

The dosage typically ranges from about 0.3–1.0% depending on manufacturer and onthe degree of washout resistance required. It is generally preferable to use the lowest dosethat is consistent with the degree of washout resistance needed. The anti-washout admixturesare usually powder-based products, which makes dispensing difficult. Liquid productsare available but may be less effective or also have storage/dispensing problems.

4.9.1 Mechanism

The anti-washout admixtures produce an open-branched polymer network in the water,which locks together the mix water, cement and sand, reducing the tendency for dilutionwith external water during and after placing. The cohesion of the mix is increased,reducing workability and flow. As underwater concrete needs to have high flow and beself-compacting, superplasticizers are needed to recover the lost workability.

4/30 Admixtures for concrete, mortar and grout

4.9.2 Use

Anti-washout admixtures work on the mix water but are more effective if the base mix iscohesive. Cement contents should be at least 400 kg/m3 but this can include blendedcement. The increased cohesion provided by very fine binders such as silica fume havebeen found particularly effective in some applications. Sand contents should also be high,typically 45% or higher, and need to have a uniform medium or fine grading.

The concrete can be placed by skip, pump or tremie and at higher admixture dosagesmay tolerate some freefall. However, care is needed not to allow the concrete to fallthrough a water-filled pump line or tremie pipe, as the turbulent flow produced will causethe mix to segregate.

There is no currently recognized British Standard to measure the performance of anti-washout admixtures. Most comparisons are based on plunge tests where a known mass ofconcrete is dropped through water a number of times in a wire cage and the degree ofwashout determined as percentage loss in mass.

4.10 Pumping aids

Modern pumps are able to cope with most concrete mixes but difficulties can be experiencedin the following situations:

• Concrete that is produced with sand that is poorly graded or coarse aggregate that iselongated or flaky. This can lead to blockages or segregation within the pipeline andcan usually be helped by a low-level air entrainment.

• Pumping for long distances or to high levels can lead to very high pump pressures andthe need for improved lubrication within the lines. Pumping aids can also providebetter workability retention for long pumping distances and can be particularly effectiveif there is likely to be delays or breaks in the pumping when segregation or stiffeningin the line could prevent a restart.

• Lightweight aggregates are often porous, with water absorption much higher thannormal aggregate. Pumping these concretes presents problems, particularly if theaggregates are not pre-soaked and fully saturated. The pump pressure drives water intothe aggregate causing the mix to dry out and block the lines. Specifically designedpumping aids are available to assist in the pumping of lightweight aggregate concretesby reducing the amount of water which becomes absorbed and allowing drier aggregateto be used.

The principal chemicals used are long-chain polymers of various types or surfactantssimilar to those used for air entrainment. The surfactants are usually blended with water-reducing admixtures, which allow slightly increased workability without loss of properties.The long-chain polymer gives cohesion to the mix and prevents bleeding and water loss.They help to maintain a layer of cement paste round the coarse aggregate particles and toproduce a slippery layer at the concrete/pipe interface. Surfactants act by increasing pastevolume and reducing aggregate interlock.

Admixtures for concrete, mortar and grout 4/31

4.11 Sprayed concrete admixtures

Sprayed concrete is pumped to the point of application and then pneumatically projectedinto place at high velocity. The applications are frequently vertical or overhead and thisrequires rapid stiffening if slumping or loss by concrete detaching from the substrateunder its own weight is to be avoided. In tunnelling applications, sprayed concrete isoften used to provide early structural support and this requires early strength developmentas well as very rapid stiffening.

Admixtures can be used in the fresh concrete to give stability and hydration controlprior to spraying. Then by addition of an accelerating admixture at the spray nozzle, therheology and setting of the concrete are controlled to ensure a satisfactory build-up on thesubstrate with a minimum of unbonded material causing rebound.

There are two spraying processes:

• The dry process where the mix water and an accelerator are added to a dry mortar mixat the spray nozzle.

• The wet process where the mortar or concrete is pre-mixed with a stabilizer/retarderprior to pumping to the nozzle where a liquid accelerator is added.

The wet process has become the method of choice in recent times as it minimizes dustemissions and gives more controlled and consistent concrete.

Accelerators are of two types:

• Alkaline, which give the quickest set, often less than 1 minute, and very early strengthdevelopment but presents a safety hazard to the applicators and tend to give low laterage strengths.

• Alkali-free, which can take tens of minutes to set but is safer for the applicator and haslittle effect on later age strength.

Sprayed concrete admixtures are often affected by the cement chemistry and their userequires technical and practical expertise. For these reasons, their use should be left tospecialist applicators.

4.12 Foamed concrete and CLSM

Low-strength, low-density concrete has a number of applications including backfilling ofutility trenches, filling old tunnels, sewers, tanks and other voids in or beneath structures.CLSM, controlled low-strength material is a sand-rich concrete with a high dose of apowerful air-entraining admixture. This gives air contents up to about 25% with densitiesdown to about 1800 kg/m3 and strengths of about 4 N/mm2. To get lower densities, the airmust be added in the form of a foam, which is produced from a surfactant admixture ina foam generator. This is then blended with a sand cement mortar or with a cement groutto give densities down to 600 kg/m3 or lower. Strength depends on the cement contentand density but is typically in the range 0.5–10 N/mm2.

The foamed concrete/CLSM is very fluid and essentially self-levelling. It can flowlong distances and completely fill difficult shaped voids. It is free of bleed or settlementand as it sets, the heat of hydration expands the air, tightening the mortar into the void so

4/32 Admixtures for concrete, mortar and grout

that no gaps remain. Its low strength and high level of air entrainment make it easy toremove at some future time. It has many other advantages related to specific applications.

4.13 Other concrete admixtures

4.13.1 Polymer dispersions

Polymer dispersion admixtures are aqueous dispersions of elastomeric/thermoplasticsynthetic polymers that will form a continuous film under the conditions of use whensufficient water is lost from the system.

Typical dose rate is 10–30 litres of a 50% solids dispersion per 100 kg of cement.They:

• Reduce the permeability of concrete, improving durability.• Give a more reliable bond to base concrete and to reinforcing steel.• Increase flexibility and toughness and resistance to abrasion/dusting.

For economic reasons applications are usually restricted to thin section concrete, e.g.concrete up to about 75 mm in thickness. Polymer dispersions are also widely used inscreeds, mortars, renders, and grouts.

4.13.2 Pre-cast, semi-dry admixtures

Admixtures are used in the manufacture of ‘dry’ or ‘semi-dry’ vibrated and pressedconcrete products such as paving and masonry blocks, bricks, flags and architecturalmasonry to assist compaction, disperse cement, colour and fine aggregate particles, andhelp to control primary and secondary lime staining (efflorescence). Admixtures can bespecific to only one of these functions or can be multi-purpose.

Semi-dry concrete mixes with very low water contents (typically 6–9%) are madeeasier to compact by admixtures that reduce particle attraction and surface tension, makingthe mix more susceptible to vibration and pressure, thereby increasing compacted density.The reduced particle attraction prevents agglomeration and disperses cement and colourparticles.

The plasticizing action of these admixtures is only effective under vibration and mustnot lead to an increase in workability as this could cause the mix to deform before setting.

4.13.3 Truck washwater admixtures

Washwater admixtures are designed to overcome the problem of disposal of washwaterfollowing the cleaning of the inside of truck mixer drums at the end of the day. Thesolution to this is to recycle the washwater into the following day’s concrete by the useof a truck washwater admixture.

The washwater admixture stabilizes the concrete washwater in the drum of a readymixedtruck on an overnight or over-weekend basis, preventing the cement from hydrating andsolid material from forming a layer of hard settlement at the bottom of the drum or on the

Admixtures for concrete, mortar and grout 4/33

blades. The washings are then incorporated as water and filler in the production of thenext day’s initial load.

Dosage varies between manufacturers depending on the materials used and the periodbefore fresh concrete is to be mixed but is typically between 1 and 3 litres for a 6 m3

truck. The driver operates a lance that discharges the admixture and water at high pressureto wash the residue to the bottom of the drum. The dispenser is programmed to deliverexactly 200 litres of water into the drum. The drum is rotated to finally wash off theresidues and is then left static until the next day (or over a weekend if appropriatelydosed). The only alteration to the next mix is to reduce the mix water by 200 litres.

4.14 Mortar admixtures

Although any combination of sand and cement can be thought of as a mortar, mortaradmixtures are usually intended for use in mixes of 3:1 or leaner for use in bedding ofmasonry or as renders. At this sand content, the plasticizing admixtures used for concreteare relatively ineffective but air entrainment lubricates the interlock between sand particlesgiving large reductions in water content. For this reason, air entrainers are sold as andusually called mortar plasticizers.

Almost any surfactant, including washing-up liquid, will have this effect but controllingthe air content for different types of mix, mixer, mixing times, and then providing stabilityagainst air loss up to the time of hardening is not so simple. The admixtures supplied areoften multi-component blends of surfactants and stabilizers designed to optimize theserequirements.

A good plasticizer provides additional benefits by reducing density, improving thetrowelling characteristics and providing freeze–thaw resistance from soon after the mixhas hardened.

Masonry mortar is often provided by readymix in the form of a ‘Ready-to-Use Mortar’,which has been retarded to allow use for typically up to 36 hours after delivery. Whenplaced between absorbent masonry units, some water and retarder are sucked out and thisinitiates the hardening process of the mortar between the units even though the unusedpart of the mortar delivery is still plastic and usable. If the masonry is not absorbent,‘Ready-to-Use Mortar’ should be used be used with care and retardation restricted to 8hours’ working life. In the case of renders, working life should always be restricted to 8hours’ working life or a soft, dusting surface is likely to be obtained. Air entrainers for‘Ready-to-Use Mortar’ are particularly critical, as any air loss over the 36-hour workinglife will result in a serious loss of consistency.

Other admixtures can be used with mortars, the most common being:

• Waterproofing admixtures to reduce moisture movement through renders.• Polymer dispersions, to aid bonding, increase flexibility and give some water repellency.• Water-retaining admixtures to reduce the suction of water from the mortar into the

masonry unit and enhance cohesion/reduce bleed.

Only mortar plasticizers and retarders are currently covered by standards BS 4887 andEN 934-3.

4/34 Admixtures for concrete, mortar and grout

4.15 Grout admixtures

Grout normally refers to a fluid mix of cement and water but may also include sand upto about 2:1. Most conventional concrete admixtures are effective in these mixes, especiallysuperplasticizers, which can give large water reductions while retaining a very fluid mix.Air entrainers are not very efficient in cement-rich mixes.

EN 934-4 refers to grout admixtures mainly with respect to grouts for filling the ductsin post-tensioned structures. The standard requires high fluidity at a w:c below 0.42,fluidity retention reduced bleed and shrinkage while the grout is still fluid and also coversthe use of expanding admixtures to overcome early age volume loss. These admixturesare normally gas-expanding systems that release hydrogen or nitrogen over a periodwhile the grout is fluid. This bulks out the grout, helping it to expand and fill any voidscaused by settlement. Bleed water can be forced out of a venting valve at the high pointof the duct. The stress induced by these admixtures is limited because of the compressibilityof the gas. Expansive cement should not be used for this application as the volumeexpansion occurs after the grout has hardened, which is too late to be effective andcreates high stresses that can crack the concrete element.

Anti-bleed admixtures are also available. These are mainly used in tall vertical ductswhere water separation can result in the formation of large water lenses. The problem isparticularly bad when 7-wire cables are used as water can wick up the inside of thestrands. The anti-bleed admixtures not only prevent cement settlement but also seal thegaps between wires in the strands, reducing the wick effect.

4.16 Admixture supply

4.16.1 Suppliers

Admixtures should only be accepted from a reputable company able to demonstrate thatthey operate a third-party certified quality management system complying with ISO 9001or 9002. The company must be able to show that their products have been properly testedand comply with the appropriate standard where this is applicable.

Suppliers should be able to provide a data sheet outlining the properties and performanceof the product. It should give dosage information including the effects of overdosing, anysecondary effects such as retardation or air entrainment, instructions for storage, handlingand use, compatibility with other admixtures and cement types, and any other informationon situations where there may be restrictions or special requirements on use. The appearance,specific gravity, solids, chloride and alkali contents of the admixture should be clearlystated. A sheet covering health and safety data (MSDS) must be provided for the purchaserand the user at the batching plant. The supplier should also provide a reasonable level oflocal technical backup for the products.

In the United Kingdom, the CAA (Cement Admixtures Association) requires that allthe provisions detailed above are met by its members and in Europe, national associationswho are members of EFCA (European Federation of Concrete Admixture Associations)also meet these requirements.

Admixtures for concrete, mortar and grout 4/35

4.16.2 Storage

Admixtures are usually stored in bulk, in plastic or metal tanks or in barrels. Modern bulkplastic storage tanks are double-skinned in order to minimize the risk of leakage due toa skin splitting. If double-skinned tanks are not used, the admixture storage area shouldbe bunded in order to prevent any spillage entering the water system.

Admixtures should be protected from freezing, all tanks and pipes should be clearlymarked with their contents and the shelf life of the product should be indicated.

Admixture containers should never be left with open tops due to the risk of contamination.Rain can also enter, diluting the product and reducing its effectivness.

If an admixture exceeds its storage life or is no longer required, it must be disposed ofthrough an approved contractor.

4.16.3 Dispensers

The dosage of most admixtures is only in the range 0.2–2.0% volume for weight oncement. This means that relatively small changes in actual volume of admixture dispensedcan have a large effect on dosage as a percentage of cement and thus significantly affectthe quality and properties of the concrete. For this reason, it is important that admixturesare added to the concrete through an accurate, frequently calibrated dispenser.

The dispenser should include an automatic water-flushing system to clean it andensure that all admixture is fully washed into the mixer. This is essential where severaladmixtures are dispensed through the same dispenser or cross-contamination can occur,potentially leading to unwanted effects.

There are three main types of admixture dispenser used in the UK:

• Volumetric measurement• Weighing• Volumetric metering

All three types of dispenser have a successful history of use but each has its own strengthsand weaknesses. The admixture supplier should be able to give advice on the mostappropriate type.

4.16.4 Time of admixture addition

For optimum performance most admixtures should be added at the end of the batchingprocess, with some of the mixing water. Liquid admixtures should never be added directlyonto dry cement. The timing of the addition affects the performance of the admixture,therefore it is important that the admixture is always added at the same point in thebatching process to ensure consistent performance. Particular attention to the mixing isneeded to ensure that all the admixture is uniformally distributed through the entire mix.

Certain admixtures can be added on-site, e.g. some superplasticizers and foamingadmixtures, and it is important that in this case accurate and safe dispensing methods areavailable. All admixture additions should be through a calibrated, metered pump and becarried out under supervision and to a prepared plan agreed by all parties.

4/36 Admixtures for concrete, mortar and grout

4.17 Health and safety

The major admixture companies all supply safety data sheets for their products and theseshould be carefully checked and any advice on handling passed to those working with theproducts. Most admixtures are water-based and are non-hazardous or, at worst, irritants.A small group have a high pH and are therefore harmful/corrosive: this particularlyapplies to some air-entrainers, waterproofers and shotcrete accelerators. A very smallnumber; mainly from the accelerators and corrosion inhibitors, may be toxic if ingested.

Some organic admixtures have a high oxygen demand under biological breakdownthat, during disposal or in the event of a major spillage, may cause problems for fish inrivers and in processing of effluent in water-treatment plants. Any such problem shouldbe notified to the appropriate authorities immediately it occurs.

Further reading

A Guide to the Selection of Admixtures for Concrete Concrete Society TR 18Cement Admixtures: Use and Application CAAChemical Admixtures for Concrete RixomApplications of Admixtures in Concrete PaillereConcrete Admixtures Handbook RamachandranSuperplasticisers & Other Chemical Admixtures CANMET/ACI Rome 97 ACI SP-173Lea’s Chemistry of Cement & Concrete HewlettWater-reducing admixtures in concrete BRE Information Paper IP 15/00

The UK Cement Admixtures Association also produces a number of guidance papers (furtherdetails on their website, admixtures.org.uk.

PART 4

Aggregates

This Page Intentionally Left Blank

5

Geology, aggregatesand classification

Alan Poole and Ian Sims

5.1 Introduction

Natural rock in the form of aggregate particles typically makes up between 70 per centand 80 per cent of the volume of a normal concrete. The word ‘aggregate’ is a familiarone, perhaps because rock fragments or gravels are one of the most important commonbulk materials in general use, particularly in the building industry. Natural sand, graveland crushed rock undoubtedly form a major and fundamental part of concrete and mortars,but they are also very important constituents of road-making asphalts and macadams, andare used extensively as filters and drainage layers, and as railway ballast.

Particles of natural rock are by far the commonest form of aggregate, but recycledcrushed concrete and manufactured materials such as furnace slags and expanded clay,shale or slate pellets are also used to a more limited extent. The aggregate as a materialmust be strong, durable and inert to give satisfactory performance, and the sizes of theconstituent particles must be appropriate for the intended application. They are describedas coarse aggregate, in the UK if particles are retained on a screen with 5 mm apertures,or 4 mm apertures, if recent European standard specifications (CEN TC154SC2, EN 12620,2002) are followed. They are described as fine aggregate, or sand if they pass throughthem. Aggregates are normally separated into size fractions by the use of a series ofdifferent sized screens, but within any fraction there will be a grading of sizes from thoseparticles that can just pass the larger screen to the ones that are just fractionally too largeto pass the smaller screen. Where a wider range of sizes is necessary for, say, a concrete,

5/4 Geology, aggregates and classification

several size fractions may be selected and mixed together to give the appropriate overallgrading.

5.2 Fundamentals

The rocks of the Earth’s crust such as those we see on the surface today were formed atvarious times in the geological past and in a variety of ways. The geological time scalehas been developed as the result of many research studies all starting from the pioneeringwork of William Smith in 1815 (Winchester, 2001). A modern consensus of the geologicaltime scale is shown in Table 5.1.

Table 5.1 The Geological Column, based on information from Eagar and Nudds (2001) and other sources

Age (Ma BP)* Epoch Period Era Eon

Holocene (recent)0.01 Quaternary

Pleistocene2.5

Pliocene5.2 Neogene Cenozoic

Miocene23.5 Tertiary

Oligocene35.5 Paleogene

Eocene56.5

Palaeocene65

Cretaceous146

Jurassic Mesozoic205

Triassic251

Permian290

Carboniferous353

Devonian Palaeozoic409

Silurian439

Ordovician510

Cambrian540

Precambrian Proterozoic2500

Precambrian Archaean4500 Formation of

the crust4600 Origin of the Earth

*MaBP: Millions years before the present.

(Num

erou

s ep

ochs

wor

ldw

ide)

Pha

nero

zoic

(‘e

vide

nt’

life

, i.e

. m

any

foss

ils

in s

edim

enta

ry r

ocks

)

Geology, aggregates and classification 5/5

The geological age of a source of rock for aggregate is usually given along with itsname. This allows the rock to be placed in its correct geological setting and history. Also,with many rocks, particularly sediments, the age may provide a clue to its properties.Although there are many exceptions, geologically older rocks are likely to be betterlithified, that is, solidified, and consequently are likely to be harder and more durablethan younger ones of the same type. As an example, the differences in the physicalproperties of Cretaceous Chalk and Carboniferous Limestone are obvious, though bothare limestones. A major factor accounting for the difference is that the chalk is about 300million years younger than the Carboniferous Limestone.

All natural rock materials are composed of aggregations of one or more mineralconstituents. A mineral is defined as a natural inorganic substance with a definite chemicalcomposition, atomic structure and physical properties (Kirkaldy, 1954). Often the individualmineral crystals or grains are very small, with some rocks the minerals are so small thatthey cannot be seen without the aid of a magnifying glass, but in others, notably certaingranites, the individual mineral crystals can be several centimetres long.

The chemical elements from which these minerals are formed are of course thosepresent in the Earth’s crust. Perhaps surprisingly, eight chemical elements account formore than 99 per cent by mass of all crustal rocks as is shown in Table 5.2. These eightelements combine together in various ways to form the common rock minerals and maybe summarized in tabular form as in Table 5.3.

Table 5.2 The abundance of chemical elements in theEarth’s crustal rocks by mass. (After Carmichael,1982)

Element % mass

Oxygen 46.4Silicon 28.2Aluminium 8.2Iron 5.6Calcium 4.2Magnesium 2.3Sodium 2.4Potassium 2.1Others 0.6Total 100

Table 5.3 The common rock-forming minerals

Minerals RD* H**

Silica Quartz SiO2 2.65 7

SilicatesPrimary Felsic (light) Al, Ca, Na, K, Si 2.57–2.74 6.65

Mafic (dark) Mg, Fe, Si 2.8–4.0 5–5.7Secondary Clay minerals Al, Si, K 2.2–2.4 2–3

Carbonates Calcite CaCO3 2.71 3

* RD The relative density range of the minerals** H This is the typical hardness of the mineral(s) based on the Mohs’ Scale of mineral scratch hardness wheretalc is given a hardness of 1, quartz 7 and diamond 10. (The fingernail will scratch up to 21/2 and a penknifeblade 61/2)

5/6 Geology, aggregates and classification

Rocks may be divided into three broad groups. The first, igneous rocks, are formedwhen molten rock material (called magma) is generated below or within the Earth’s crustand crystallizes as solid rock as it cools down, either on the surface as a lava or within theEarth’s crust as an intrusion. Since igneous rocks are intruded into pre-existing rocks invarious ways and are now seen at the Earth’s surface as a result of erosion, the series ofdescriptive terms used to describe their occurrences are best illustrated pictorially as inFigure 5.1.

Figure 5.1 An idealized block diagram showing common forms and associations of intrusive and extrusiveigneous rocks (after Fookes, 1989).

Extrusiveflows(lava)

Volcano

Neck

Sill

Dyke

Aureole

LaccolithBatholith

Tor

The second group are sedimentary rocks which are formed by the accumulation offragments of pre-existing rocks resulting from processes of erosion, organic debris suchas shell fragments or plant material. These are the detrital sedimentary rocks, or alternatively,they may be formed as a chemical precipitate from oversaturated sea, or ground waters,the chemical and biochemical sedimentary rocks.

The third group, metamorphic rocks, are formed from pre-existing rocks of any type,sedimentary or igneous, which have then been subjected to long periods of increasedtemperature and /or pressure within the crust. Depending on the severity and the timerocks are subjected to these high temperatures and pressures they undergo progressivechanges (‘metamorphism’) resulting in new minerals being formed and modifications totheir original appearance (McLean and Gribble, 1990).

5.3 Geological classification of rocks

Rocks are usually identified by their rock name and sometimes also the geological periodin which they were formed. Geologists have devised numerous rock name classificationsusually based on the sizes and types of mineral particles or crystals that they contain.Some are very elaborate and detailed, but for general purposes and as a first step inaggregate identification, simple systems for igneous, sedimentary and metamorphic rockssuch as are illustrated in Tables 5.4–5.6 should be adequate.

It was noted above that metamorphic rocks are formed from pre-existing rocks by theaction of heat and /or pressure extending over a variable period of geological time. Theseare consequently difficult to classify in detail because of their variety and the range of

Geology, aggregates and classification 5/7

Table 5.4 Simplified igneous rock classification

CompositionOccurrence

Acid Intermediate Basic Ultrabasic

Silica% >65 55–65 45–55 <45Quartz Much quartz Little quartz Almost no No quartzcontent quartzColour Pale Dark Very darkDensity 2.4–2.7 2.7–2.8 2.9–3.0 3.0

Coarse- Granite Diorite Gabbro Peridotite Large intrusionsgrained(>2 mm)Medium- Micro- Micro-diorite Dolerite Small intrusionsgrained granite Sills(2–0.5 mm) DykesFine- Rhyolite Andesite Basalt Lava flows andgrained small dykes(<0.5 mm)

Table 5.5 Simplified sedimentary rock classification

Main mode of origin Dominant chemical Rock name group Examplesconstituents

Detrital Silica and silicates Terrigenous ConglomerateSandstoneMudstoneShale

Chemical and Carbonates Carbonate rocks Limestonebiochemical Chalk

Chlorides and sulfates Evaporites Rock salt, GypsumVarious forms of silica Siliceous rocks Chert, FlintFerruginous minerals Ironstones and iron-rich Sedimentary

rocks ironstonesPhosphates Phosphorites Phosphorites

Biological Carbon and organic Carbonaceous rocks Coalresidues Sapropelites

Deposits from Silicates Volcaniclastic rocks Agglomeratesvolcanic sources Tuffites

Ignimbrites

Table 5.6 Simplified metamorphic rock classification

Regional metamorphic rock groups

Low grade Moderate grade High grade

Grain size <<1 mm 1–3 mm >3 mm

Temperature <300°C 300–550°C >550°C

Pressure <3 kb 3–5 kb >5 kb

Rock fabric Slaty cleavage Foliation Discontinuous banding

Rock types Slates Schists of all types GneissesPhyllites Migmatites

Granulites

Gra

in s

ize

5/8 Geology, aggregates and classification

possible metamorphic conditions. Table 5.6 divides the general classification into three,but if the effects of the high metamorphic temperatures and pressures are simplified totwo, namely, high grade and low grade (effects often associated with how deeply therocks are buried in the Earth’s crust, since typically both temperature and pressure risewith depth) and related to the original rock types, then a much simplified classificationappropriate to possible aggregate materials can be devised as indicated by Table 5.7.

Table 5.7 A generalized classification of metamorphic rocks with special reference to rocks potentially suitablefor aggregate

Metamorphic environment – temperature and pressure conditionsOriginal rock

Low grade – shallow burial Medium and high grade – deep burial

Foliated platy rocks(effects of pressure paramount)

Clay, shale and volcanic slate phyllite, schist gneiss*tuff clayey sandstone greywacke-sandstone*† quartz-mica schist,

fine-grained gneissslaty marble*† calcerous schist

Clayey limestone granite,* shered granite, slate granite-gneiss*, quartz-micaquartzitic tuff schist

Basalt,* basic volcanic green (chlorite) schist amphibolite (amphiboletuff schist) hornblende gneiss*

Non-foliated, massive rocks(effects of increased temperatureparamount)

Any parent rock hornfels* (somerecrystallization but oftenoriginal features still present)

Quartzose quartzitic sandstone* quartzite (includes mica ifsandstone* impure) psammite*,

granulite* (essentially quartz,feldspar + some mica)

Limestone* and marble* (if impure, contains adolomite* wide range of calcium and

magnesium silicates withincreasing metamorphic grade)

*Rocks most likely to be useful for aggregate†Some rocks containing clay minerals may not perform satisfactorily as aggregate in some applications

The classification of the most common sedimentary rocks, that is, the clastic (ordetrital) rocks and the limestones, also require further elaboration. Clastic sediments arenamed on the basis of particle grain size and whether they are unconsolidated (sands andgravels) or have been consolidated by lithification (hardened, or indurated). This classificationis given in Table 5.8.

Limestone, although essentially simple, in that all limestones are composed mainly ofcalcium carbonate (CaCO3) prove difficult to classify in detail, because when impure theycan fall into a range of sedimentary compositions, with a claystone or sandstone at oneextreme of the classification, and pure limestone at the other, as is illustrated in Table 5.9.

A limestone can also contain a proportion of dolomite, which is a calcium–magnesiumcarbonate mineral (CaMg(CO3)2). If the rock contains over 90 per cent of mineral dolomiteit is called a dolomite (named from the locality where it is found in the Tyrolean Alps),if there is more than 50 per cent dolomite it is calcitic dolomite, while less than 50 percent, it becomes dolomitic limestone.

Geology, aggregates and classification 5/9

Yet another complication arises in that many limestones are composed in part fromshell fragments and other clastic particles of calcium carbonate. These are called allochemicalconstituents of a limestone and are listed in Table 5.10. Limestones may contain one ormore of these constituents set in a lithified lime mud (micrite), or a more coarse recrystallizedcalcite (sparite).

Table 5.8 The classification of detrital (clastic) sedimentary rocks based on grain size

Grain diameter, Category Term Indurated equivalent Ajectival termsmm

>2.0 – Gravel Conglomerate Rudaceous0.60–2.00 Coarse Sand Sandstone0.20–0.60 Medium0.06–0.20 Fine

Arenaceous.020–.060 Coarse Silt Siltstone.006–.020 Medium.002–.006 Fine<0.002 – Clay Mud-rock Argillaceous

Table 5.10 The allochemical constituent particles commonly found in limestones. They can be used as prefixesto refine the Folk (1959) classification given in Table 5.11

Component Prefix for use in Table 5.11 Mode of origin

Fossils bio- Skeletal parts of carbonate-secreting organisms

Peloids pel- Spherical or angular grains without internal structure,mostly of faecal origin

Ooids oo- Spherical, or sub-spherical grains with a concentricinternal structure, probably of biochemical origin

Aggregates clastic- Aggregates of several carbonate and other particlescemented together

Intraclasts intra- Fragment of previously lithified, or partly lithified,sediment incorporated into the new sediment

All surface rocks degrade with time under the effects of the weather, eventually changingto a soil, although some are much more durable than others. Research into the weatheringof rock has resulted in a generally accepted sixfold scale of rock weathering, with freshrock as 1 and soil at 6. This scale is shown in Figure 5.2.

Table 5.9 A simple classification of carbonate rocks. (After Fookes and Higginbottom, 1975)

0Claystone Siltstone Sandstone Conglomerate 0

5 10Marly claystone

20Limey marlstone Calcerous siltstone Calcerous sandstone Calcerous conglomerate

35 50Marlstone

65Clayey marlstone Silty limestone Sandy limestone Conglomeratic limestone

80 90Marly limestone

95Limestone Limestone 100

100%P

ercentage carbonate

Percentage carbonate

Table 5.11 A classification for limestones after Folk (1959) appropriate for shelly limestones. The prefixes from Table 5.10 are appropriate for other allochems

Low water turbulence High water turbulence

Subequal spar&

lime mud

Over 23

spar cement

(Sparry calcite)

Over 23

lime mud matrix

(Micrite)

Representativerock terms

PercentAllochems‘grains’

Terrigenousanalogues

Lime mud matrix Sparry calcite cement

Micrite anddismicrite

Fossili-ferrousmicrite

Packedbiomicrite

Poorly washedbiosparite Unsorted

biosparite

Sortingpoor

Submaturesandstone

Sortedbiosparite

Sortinggood

Maturesandstone

Roundedbiosparite

Round andabraded

Supermaturesandstone

Sparsebiomicrite

0–1% 1–10% 10–50% Over 50%

ClaystoneSandy Clayey or

claystone immature sandstone

Humus/topsoil

VIResidualsoil

VCompletelyweathered

IV

Highlyweathered

IIIModeratelyweathered

IISlightlyweathered

IB Faintlyweathered

IA Fresh

A. Idealized weathering profiles –without corestones (left) andwith corestones (right)

Rock decomposed to soil

Weathered/disintegrated rock

Rock discoloured by weathering

B. Example of acomplex profilewith corestones

All rock material converted to soil: mass structureand material fabric destroyed.Significant change in volume

All rock material decomposed and /ordisintegrated to soil.Original mass structure still largely intact

More than 50% of rock material decomposedand/or disintegrated to soil.Fresh/discoloured rock present as discontinuousframework or corestones

Less than 50% of rock material decomposedand/or disintegrated to soil.Fresh/discoloured rock present as continuousframework or corestones

Discoloration indicates weathering of rockmaterial and discontinuity surfaces.All rock material may be discoloured by weatheringand may be weaker than in its fresh condition

Discoloration on major discontinuity surface

No visible sign of rock material weathering

Figure 5.2 A pictorial representation of weathering grades. (After Geological Society Engineering Group Working Party Report, 1995.)

5/12 Geology, aggregates and classification

As an example, the changes in the physical properties that occur as a granite is increasinglyweathered are shown in Table 5.12. This table gives the physical properties in terms of aseries of test values appropriate to each of the weathering grades. A similar sequence ofchange occurs with other rock types as they become weathered. Clearly rock aggregatesfor concrete need to be both strong and durable. This implies that a rock suitable for useas a concrete aggregate will normally need to be in weathering grades one or two.

5.4 Sources and types of aggregates

Natural rock, sands and gravels are by far the commonest source of aggregate worldwide.Artificial and recycled materials account for only a tiny fraction of the total aggregateproduced. In the UK alone, the total aggregate requirement is in excess of 200 milliontonnes annually, while in Japan the figure is 420 million tonnes (Fox, 2002).

The production of aggregates from all sources in recent years for the UK is illustratedin graphical form in Figure 5.3 and shows that land-based extraction of sand and gravelaggregate together with crushed rock are the predominant sources, sea-dredged materialsaccount for about 10 per cent of the total while the contribution of manufactured andrecycled materials is currently very small. In recent years the demand for aggregate in theUK has not increased, but generally demand closely follows national economic trends.

In the UK, as elsewhere in the world, the highest usage of aggregates are in areas ofhigh population density and where major constructional development is taking place.Hence, in Britain demand for aggregate typically tends to be heaviest in the south-east ofEngland and in areas of the Midlands.

Sources of aggregates fall into the broad categories of sands and gravels and crushedrock. Sands and gravels are the products of erosion of pre-existing rocks and are usuallytransported by water, or by ice in formerly glaciated regions. They are typically depositedas relatively thin layers at the foot of mountains, in river valleys or along shorelines.Crushed rock, by contrast, is obtained from rock quarries which imply that appropriaterock must occur at the Earth’s surface where a quarry can be developed.

The geology of the British Isles is such that if a line is drawn southwards from theHumber down to Portland Bill, then as a generalization no strong rocks suitable forquarrying for aggregate occur to the east of the line, while to the north and west of it thereare many areas where rock quarries could be sited, or have already been developed (theStone Line, Figure 5.4). As a consequence the aggregate requirements for the denselypopulated south-east must be met by local sand and gravel deposits, by imported materialsor by transporting aggregates long distances from aggregate quarries in the west. Long-haul distances for aggregate are generally undesirable for both environmental and economicreasons (the cost of aggregate can be doubled for every 10–15 km of haul distance).

A more detailed examination of a geological map of the British Isles allows generalizationsto be drawn concerning the location and types of hard rock suitable for aggregate. Similarly,consideration of the present and past river systems, the extent of past glaciations acrossBritain and the extent of the shallow continental shelf, again allows inferences to bedrawn concerning sand and gravel resources available within the British Isles. Thesegeneralizations are illustrated on the maps in Figure 5.4.

Aggregate is produced from quarries by drilling and blasting the quarry face withexplosives to maximize the fragmentation, while minimizing the production of ‘fly rock’

Table 5.12 The typical physical properties of granite in the different weathering grades. (After Fookes et al., 1971.)

Mass Uniaxial Bulk Saturation Aggregate Aggregate Aggregate MgSO4 Secondaryweathering compressive density moisture impact impact abrasion soundness mineralsgrade strength saturated content value value value value (%)

saturated (g/cm3) (%) (%) modified (%) (%)(MN/m3) (%) retained

I Fresh 262 2.61 0.11 6 7 3.5 100 6

Stained rim 232 2.62 0.15of block

8 10 4.7 99.9 10Whole sample 136 2.58 1.09II 90% stained

II Completely 105 2.56 1.52 14 16 8.0 99.8 12stained II block

Rock core of 46 2.55 1.97III block

III–IV 24 49 17.1 66.6 17Rock core of 26 2.44 4.13IV block

WeaklyV cemented soil 5 2.24 10.00 nd nd nd nd

(coreable)

5/14 Geology, aggregates and classification

and vibration. The blasted fragments are then crushed and passed through a series ofvibrating screens to produce the particle size fractions required. These sizes depend onlocal requirements, but a typical example using the European aggregate standardspecifications (CEN TC154SC2, EN 12620, 2002) might be 20–14 mm, 14–12.5 mm,12.5–6.3 mm for coarse aggregate, and 4–2 mm and 2–1 mm for the fine. Usually therock fragments are reduced to required sizes in a series of steps with reduction ratios ofbetween 10:1 and 5:1. Usually secondary and even tertiary crushing and screening sequencesare required with oversize material being recycled through the system and the unusablewaste fine dusts removed. Many rock types are suitable for the production of a durableaggregate but the shape of the particles is also an important factor. They ideally should be‘equant’, that is, equidimensional in shape. The natural joints in the rock, together withtheir physical properties, may tend to produce elongate or flaky particles, so that carefulchoices of crushing plant design must be made in order to counter this tendency.

The aggregates won from natural sand and gravel deposits have several advantagesover crushed rock aggregate. Nature, through the processes of erosion and deposition,

350

300

250

200

150

100

50

Mill

ion

tonn

es

1965 1970 1975 1980 1985 1990 1995 2000

Crushed rock Sand and gravel Total

Figure 5.3 The production of aggregate from various sources from the UK over the period 1965–1998.(Sources: Quarry Products Association and BGS; Smith and Collis, 2001.)

Geology, aggregates and classification

5/15

Figure 5.4 Generalized maps of the British Isles showing the areas of sands and gravels, and of rocks which may be suitable as aggregates.

Aberdeen

(a)

Edinburgh

Newcastle

The stoneline

Manchester

N

Marine sandand gravellicensed areas

Maximum of theAnglianglaciationMain rivervalley sourcesof sands andgravels

Birmingham

BristolLondon

Southampton Calais

Nottingham

Glasgow

Principal carbonaterock areas

Chalk

Oolitic limestones

Compact Palaeozoiclimestones

Magnesium limestones

Principal igneous rockareas

Maximumof Anglianglaciatum

5/16 Geology, aggregates and classification

tends to concentrate hard durable particles, while at the same time the weaker non-durable material tends to be broken down further to silt and clay which is transportedmore readily by flowing water than the heavier coarser particles. The processes of depositionare such that the material tends to be at least partially sorted by size and relative density,while abrasion during transportation tends to round the particles to shapes which tendtowards spherical. Rounded particles make for a more workable concrete than very angularcrushed rock fragments which may be described as giving a ‘harsh’ mix for concrete andmay make its compaction difficult. However, because sand and gravel deposits are usuallybuilt up over a period of geological time both the sources from which the materials arederived and the energy of the erosive processes may change on a seasonal, or on a longer-term basis as erosion, or tectonic uplift alters the source area. Examples might be modificationto the rainfall pattern, or changes to the river catchment area over time. Such changes willlead to changes in the quantities, sizes and distribution of the sand and gravel particlesdeposited. Consequently most gravel resources are characterized by rapid and local variationsin size distributions and degree of sorting with layers and lenses of coarser and finermaterials, which may be interleaved with each other on a scale of a few metres such asis shown in Figure 5.5. The proportion of silt and clay in a deposit which is currentlyconsidered waste and must be discarded, together with the proportion of over-size materialwhich would require crushing to reduce its size before it can be used as aggregate, areimportant economic factors to be considered in the development of a sand and gravelresource.

Another advantage of sand and gravel reserves is that they are usually loose unconsolidatedmaterials which can be dug directly with a face shovel or a similar mechanical excavator.The raw material then only has to be washed to remove clay and silt and screened to theappropriate aggregate size fractions. If oversize particles are present and are crusheddown to aggregate sizes the crushed particles are usually blended in with the natural morerounded gravel to avoid the aggregate being too ‘harsh’.

Sources of sand and gravel aggregate may be grouped under four headings as indicatedin Table 5.13.

It should of course be recognized that although many sand and gravel deposits are ofrelatively recent geological origin, ancient river systems, glacial deposits and near-shoredeposits of earlier geological periods are sometimes worked as sources of sand and gravel.

5.4.1 Temperate fluvial environments

River terrace deposits are perhaps the commonest of the alluvial deposits. These materialsare eroded from the upper reaches of rivers, carried downstream by the river, particularlyat times of flood when it is most energetic, and deposited in the lower reaches typicallyon a floodplain where the rate of flow is checked and the river can no longer transport itsburden of sediment. This sedimentary material builds up in thickness over time. Mostrivers have a history that will extend back through the Quaternary Period with its reducedsea levels during the glacial periods, and often well into the Tertiary Period. Such riverswill have had periods of rapid cutting down through the earlier alluvial deposits to adjustto reduced sea levels while at other times at the end of glaciations they would have carriedmuch larger volumes of water than is the case today and have been capable of transportinggreater sedimentary loads and larger particles than at present. The idealized block diagram

Geology, aggregates and classification 5/17

Figure 5.5 A photograph of the face of a sand and gravel pit showing lenses and layers of sand andgravel particles of different sizes. (Aller gravels, Devon.)

and cross-section in Figure 5.6 illustrates the way in which such processes give rise toriver terrace deposits. It should perhaps be noted the deposits of this type are typicallyonly a few metres thick but often are of wide geographical extent.

5.4.2 Glacial and periglacial regions

Glacial deposits are formed principally by the processes of physical erosion by movingice sheets and glaciers and are deposited at the melting ice fronts. They are found in manyparts of the world since glaciations have occurred at various times in the geological past.In Britain most of these deposits were formed during the ice ages of the last four millionyears. The material carried on and in the ice is dumped as ice melts during the recession

5/18 Geology, aggregates and classification

of ice sheets and glaciers in periods of warming between ice ages. The block diagram inFigure 5.7 shows how material is eroded by glaciers and left behind as a glacier recedesduring a period of warming. The types of material transported by ice vary considerably,material falling on to and into the ice will tend to be ill-sorted, angular and will cover awide size range from large boulders downwards to rock flour or clay. This material isreferred to as till or boulder clay. By contrast, materials carried by meltwater across,through and away from the ice front will have characteristics more similar to those offluvial or river deposits and are usually referred to as fluvio-glacial deposits. As a resultof the recent ice ages of the Quaternary much of the British Isles, north of the southernlimit of glaciation (Figure 5.4), together with large areas of Russia, Canada and parts ofthe USA, have suffered the effects of glaciations and retain glacial deposits in valleys andother regions where they have been protected from removal by more recent erosion.

Table 5.13 Types of unconsolidated sand and gravel sources

Alluvial deposits • Typically well sorted in particle size• Relatively clean• Rounded particle shape

Glacial deposits • Typically variable in shape and particle size distribution• High content of silt and clay• Angular particle shape• Higher proportion of weak and weathered particles• Fluvio-glacial outwash deposits usually the better aggregates

Coastal deposits • Rarely used, principally for environmental reasons

Marine deposits • Marine sands and gravels originally laid down as alluvial or fluvio-glacial deposits

Sands & gravels Floodplain Paired terracesRiver downcuttingthrough floodplain

Currentalluvium

Terracegravels

Currentfloodplain

Figure 5.6 The development of river terrace deposits as a river cuts down with time. (After Leet et al.,1982.)

Geology, aggregates and classification 5/19

Calculations indicate that the sea level was lowered in Quaternary times by up to150 m worldwide during the most extreme of the glaciations. Consequently rivers cutacross what is now the shallow seabed around the coasts depositing sand and gravelswhich are now being extracted by marine dredging. Materials carried from land byicebergs and sea ice during the ice ages have also contributed to these deposits. Currentareas where marine dredging is carried out are indicated in Figure 5.4. The depth limitson dredging with current equipment is about 40 m but as technology advances it will soonbe feasible to extend this limit to 50 m. However, the land transport distances fromsuitable quays, the technical difficulties and the numerous environmental constraints allplace limits on the economic viability of these resources.

As has been described above, in temperate climatic regions such as that of the BritishIsles the location of suitable aggregate resources depends either on finding localitieswhere there is suitable rock for quarrying, or on finding fluvial deposits of sand andgravel. In general terms the upland mountainous regions provide possibilities for quarrydevelopment and the production of crushed rock aggregate. While sand and gravel resourcesare usually located in the mature floodplain reaches of river valleys and in shallow coastalwaters. These types of topographic expression, and the aggregate resources they might beexpected to contain will be broadly similar and applicable in other climatic regions of theEarth, provided the effect of the particular climate on the weathering of rock and theprocesses of erosion are taken into account.

Figure 5.7 An idealized diagram of glacial and periglacial terrain during glacial retreat. (After Fookes et al.,1975 and other sources.)

Unsorteddumpedcrevasse

filling

Esker (sand,gravel, siltand clay)

Scree andalluvial fans

Sediment infilledMarginal lake(sand and silt)

Surfacestreams

Fluvio-glacialoutwashplain

Ice front

Kame(sand and gravel)

Fluvio-glacial (sand and gravelwith silt and clay lenses)

Angular ill-sortedablation till

Abrasion between blocksand with bedrock forms

comminution anddeformation till

Lodgement till (widesize distributionboulders to clay,

dense and compact)

Angular ill-sortedenglacial debris

Debris infilledcrevasse

Angular, ill-sortedsuperglacial

debris

5/20 Geology, aggregates and classification

5.4.3 Hot desert regions

In hot arid climates the processes of erosion are modified in that the breakdown of rockis controlled principally by physical processes such as expansion and contraction due totemperature cycling. The role of water in breaking down a rock is much more limited,though the rare flash floods following rainstorms do provide the energetic mechanism fortransporting rock debris from the mountainous regions to be spread as sand and graveloutwash fans at the foot of the mountains, and as alluvial plain silts, sands and gravels onthe lower ground. The lack of vegetation and the windy dry conditions favour thedevelopment of sand dunes which are often developed on the alluvial plains. The idealizedblock diagram in Figure 5.8 illustrates these various features, and the kinds of materialstypically available in hot desert regions. The high water table level, which is often saline,is also a common feature in the low ground furthest from the mountains, and introducesthe additional difficulty of salt contamination of potential aggregate resources.

Figure 5.8 An idealized block diagram of a hot desert region showing the landforms and associatedmaterials in the different zones. (After Douglas, 1986.)

4

3

2

1

Mountains (solid rock maybe suitable for quarrying)

Debris and scree fans (angular,boulders, cobbles and gravels in

an earthy matrix)

Intermittent watercourses (sands,gravels, silts and

clay – wadi gravels)

Alluvial flood sheetplain or bajada(clays, silts, sandsand gravels, gravelpavementspossible)

Ancientrocks

Unconformity

Desert deposits

Advancing barchansand dunes (single-sized, fine-grained

rounded sand grains)

1 Mountain zonehard rock quarries insuitable rock types

2 Pediment zone scree andgravel fans, wadi deposits

3 Alluvial plain-zoneAlluvial sediments, some wadi gravels, saltcontamination possible

4 Playa basin/coastal zoneDune sands, fine-grained sediments,duricrusts and salt contamination possible

The sand dunes may at first sight be thought of as an excellent source of fine aggregatebut this is usually not the case because the wind-transported sand particles tend to be veryfine-grained, single-sized and poorly graded. Sand resources in the low ground also needto be viewed with caution because salts crystallized within the sand and derived from thesaline water table may be present in sufficiently high concentrations to render themunsuitable for concrete aggregate.

Capillary rise of saline ground water can give rise to a salt-cemented layer at the landsurface, because the water evaporates leaving the salt as an intergranular cement. Suchlayers are called duricrusts, and may be a metre or so in thickness. If such a duricrust iscemented by carbonate then it may after crushing provide a valuable source of coarse

Geology, aggregates and classification 5/21

aggregate, but it will be entirely unsuitable as concrete aggregate if the cementing materialis halite (NaCl), or gypsum (CaSO4·2H2O) or if the carbonate is contaminated by suchminerals.

5.4.4 Tropical hot wet environments

Hot wet tropical climates also introduce differences in the sequence of rock weatheringand erosion which in turn is reflected in the types of material produced and the possiblesources of aggregates for concrete. The importance of water and elevated temperature onthe erosion and weathering in wet tropical conditions is critical in that the zone of rockweathering can develop rapidly, variably and is often very thick. A thickness of 60–100 mof degraded weathered rock is not uncommon and is an important consideration forquarry development. Vegetation also grows quickly adding peat, organic material andhumic acids to the soil and ground water so that mature river valleys and alluvial areas areunlikely to yield satisfactory sand and gravel resources. The moist conditions often tendto assist the chemical leaching of rock during the weathering process and lead to dissolutionof silica and calcium and leave behind higher concentrations of alumina and iron oxidesand hydroxides which may in turn form hard layers within the soil. These layers arecalled ferrocrete, bauxite and alucrete, depending on their composition. Soluble silicaand carbonates may be reprecipitated elsewhere to form silcrete or calcrete layers. Suchhard layers are often exhumed by later erosion to form duricrust layers which can be upto 10 metres thick, they may cap flat-topped hills and form a valuable local source ofaggregate. Nevertheless as can be seen in the block diagram in Figure 5.9 the possiblelocalities which may provide sources of concrete aggregates are limited in such terrains.

5.5. Classification of aggregates

The principal requirements for an aggregate to be suitable for concrete are that it will bestrong, durable and inert. The individual particles ideally must be of equant shape and beevenly graded from coarse to fine within their particular grading size fraction. Classificationof aggregates beyond the broad categories of crushed rock, sand and gravel must beappropriate to its use in the construction industry and have both a scientific and a commercialviability. A petrographic description can be of considerable value in any assessment ofthe likely performance of a particular rock type, but the geological name is only ofimportance in that it implies that there will be a set of physical and mechanical propertieswhich are very broadly typical of that particular material. In an early consideration of thepractical classification of aggregates the British Standards Institution in 1973 recognizedeleven ‘Trade Groups’ of rocks (BS 812, 1975), each group containing a number ofindividual rock types classified into their particular group on the basis of the assumedsimilarities of the physical and mechanical properties they shared. This listing is shownin Table 5.14. Although this concept of grouping rocks together according to their propertiesappeared to be useful, a number of rock types were difficult to assign to a single groupand it omitted a classification for sands and gravels. In general, the variety and range ofrock types covered was too wide, with boundaries between groups difficult to define.Also, the use of precise geological names for the groups and names used within groups

Figure 5.9 An idealized diagram of a humid tropical environment. (After Douglas, 1986.)

Sedimentarymudstone

Irrigation area on coastal awamp landaffected by acid sulphate soils

Subsurface karst cavitiescontaining limestone boulders

Pinnacles of buried karst surface

Swamp at base of karst hill

*Isolated limestone hill with dolines andpinnacles on summit. Roof collapse ofcaverns and large caves within the hill

*Colluvial footslope with boulders(often described as boulder bedsin early geological literature)

Irregular basal surface of weathering

Granite core stone boulders

Peat beneath montane forest

Valley depth obscured by variation in tree height

Small natural landslips under forest on steep slopes

Montane forest on summits

Decrease in tree height near summit

Steep valleys and sharp ridgecrests of feral relief in hills. Somesmall streams cascade overrocks, others are buried beneathforest and boulders

*Upstanding bare quartzridge

Gullied colluvial footslopewhere vegetation removed

*Isolated hill in sedimentary mudstone deeplygullied where vegetation removed

Meandering channel throughalluvial plain

Freshwater swampof acid lowlandenvironment

Mangroves at seawardedge of land

*Lenses of fine sand representing formerriver channels graded to Quaternary lowsea levels

*Potential sources of aggregates

Geology, aggregates and classification 5/23

led to unnecessary contractual disputes, so that the scheme is now omitted from thestandards.

The general consensus today is to use the correct geological name for the rock and totake careful account of its mineralogy and state of weathering or alteration. The need touse the correct geological name means that the petrographer in sampling and describingan aggregate for use in concrete must comply with the methods given in appropriateBritish Standards or their equivalents abroad. In Britain the methodologies are given inBS 812 Parts 102 (1984), 103 (1994), 104 (1994), and 105 (1989) and the Europeansimplified equivalent is BS EN 932–3 (1997). In the USA, ASTM C294–98 and C295–98 (1998) provide a nomenclature and methods for the petrographic description of aggregate.A more specific standard procedure for dealing with the particular issue of alkali–aggregatereaction (AAR) in concrete is given in BS 7943, but this particular topic is considered ingreater detail in Volume 2, Chapter 13 of this series. Reviews of special aggregate problemscan be found together with reference to the appropriate standards and guidelines in bookssuch as St John et al. (1998) and Smith and Collis (2001).

In general terms the accepted recommended approach to the description and classificationof any aggregate requires a three-stage description:

1 A description of the aggregate type (sand, gravel, crushed rock)2 A description of the physical characteristics (particle shape, size, grading)3 A petrographic description and classification (mineralogy, geological name)

To this simple aggregate classification more detailed petrological information and anytest results can be added as is necessary. This approach can be presented simply on aproforma such as is illustrated in Figure 5.10.

5.6 Aggregate quarry assessment

The periodic requirement to inspect a working aggregate quarry, and to sample theaggregates produced so that petrographic examination or other laboratory tests can beundertaken, forms an important part of quality assessment of the aggregate produced.

The most important aspect of any site inspection and sampling programme must be theessential requirement that the observations made, and the samples collected, must be bothobjective and representative. In carrying out such a programme it is first necessary to holda briefing meeting with both quarry staff and the inspection team so that the objectives ofthe inspection are clearly defined and that any other information relevant to the inspectionis made available, including information relating to health and safety matters that mayaffect the inspection procedure. A preliminary tour in the company of quarry staff isinvaluable in that attention can be drawn to features of the quarry working that may notbe obvious or have become obscured by later operations. This will result in ensuring thatno aspect of importance is missed, and consequently time will be saved, while theeffectiveness of the detailed inspection is increased. Finding a high vantage point isimportant in the early stages of the inspection since this will allow the overall geology,geomorphological and geographical setting of the site to be recognized. The past, presentand planned future working of the quarry faces need to be determined and its accuracyagreed with the quarry manager. Past, current and future predictions of production mustalso be estimated and any existing data on earlier test results need to be obtained and

5/24 Geology, aggregates and classification

evaluated against the current working situation in the quarry. Methods of sampling thequarry faces, the stockpiles of aggregate and the materials on conveyors or at stages in theproduction process are well established; for example, BS 812, Part 102 (1984) and ASTMD75–87 (1987). The overriding principle behind the procedures selected will be therequirement to obtain a representative sample so that the results of any laboratoryinvestigation or test programme can also be relied on during the period of continuedproduction.

Difficulties will arise in predicting the type and quality of future reserves beyond thecurrent existing faces. The confidence of these predictions will depend partly on the

Figure 5.10 An example of a simple proforma format for reporting the description and classification of anaggregate, after Smith and Collis, 2001. (Note: 3.1, Monomictic = single rock type, Polymictic = severaldifferent rock types.)

Classification and description of aggregate

1 Aggregate type

1.1 Crushed rock

1.2 GravelUncrushed Land won

Partly crushed1.3 Sand

CrushedMarine

2 Physical characteristics

2.1 Nominal size

2.2 Shape

2.3 Surface texture

2.4 Colour (sample condition)

2.5 Presence of fines

2.6 Presence of coatings

2.7 Extraneous material

3 Petrological classification

3.1 Monomictic Polymictic

3.2 Petrological name

3.3 Petrological description

Visual assessmentPetrographic

Quantititive analysisthin-section

3.4 Geological age

4 Full petrological description

5 Sample ref. 6 Certificate of sampling

7 Source

Geology, aggregates and classification 5/25

distance the undeveloped material is from existing worked faces and fully characterizedmaterial and partly on the potential geological variability of the source. A good geologicalevaluation of the locality and the type of material, aided by a detailed geological mappingexercise, will in most cases provide the required information to allow an adequatelyaccurate prediction of the nature and quality of the reserves. However, in materials whichshow wide local variation on a scale of a few metres, such as many unconsolidated sandand gravel deposits, or where reserves need to be ‘proved’ in detail, a scheme of trenchingor pitting will be necessary while in solid rock a pattern of boreholes will be required tosample the variation. In such circumstances the quality of the information obtained isentirely dependent on the care given to logging the pits, or the recovered drill core, andalso on the subsequent laboratory examination and testing of the samples obtained.

The technical aspects of developing a new quarry on a ‘greenfield’ site follows thegeneral pattern of investigation outlined above. In such a case the most important requirementis accurate geological information concerning both the site and the materials. This will beobtained from a desk study, supplemented by detailed engineering geological,geomorphological and hydrological mapping. Once the data have been obtained andevaluated, a programme of sampling using boreholes, pits or trenches as appropriate mustbe designed for the site together with a detailed plan for the laboratory examination andtesting of the samples obtained.

5.7 Deleterious materials in aggregates

The natural rock and gravel sources of aggregates may contain components which arepotentially deleterious when the aggregate is used in concrete and mortar. These deleteriousmaterials may be distributed throughout the entire rock or deposit, be confined to only apart of it, or alternatively be present in a particular localized feature within the sourcerock or deposit. A geological evaluation of the source combined with a thorough petrologicalinvestigation and appropriate testing will minimize the risk of deleterious materials beingincorporated into the concrete or mortar. A listing of the most common potentially deleteriousmaterials has been assembled by Sims and Brown (1998) and is presented in Table 5.14.

In addition to deleterious materials being incorporated into the aggregate from itssource, other components may find their way into an aggregate stockpile as contaminants.These may be such materials as metal, wood and plastic fragments introduced during theextraction and processing of the aggregate or as chemical solutions such as salts derivedfrom the capillary rise of saline ground waters into the base of aggregate stockpiles, orfrom salt spray or contamination from adjacent chemical storage facilities. Again, insome rare cases the processed aggregate degrades after it has been processed either in thestockpile, or in the concrete made with it. This indicates that the constituent minerals inthe aggregate particles themselves are unstable when exposed to air and water from theatmosphere.

The identification and assessment of the concentration of any of these deleteriousmaterials in an aggregate can be achieved by a combination of petrological, physical andchemical techniques. The results of such a study may indicate that further careful investigationof the source of the aggregate, its processing or its storage will be necessary in order toidentify and isolate or remove the deleterious component.

5/26 Geology, aggregates and classification

5.7.1 Interference with the setting of the concrete

Certain chemical contaminants such as soluble chlorides and sulphates, organic compoundssuch as mono- and polysaccharides, humic acid and lignins, and also soluble salts of lead,zinc, cadmium and tin, can retard the setting of cement, and hence affect the final strengthand durability of concrete. Even in low concentrations their effects are noticeable and, inextreme cases, they can prevent the proper setting of the cement binder so that theconcrete is friable or crumbly and has little or no strength. In the case of some of thesecontaminants, notably sulphates and chlorides, they remain a potential problem and threatto durability long after the concrete has set.

Chemical analysis is clearly a means of identifying such contaminants but petrographictechniques such as the selective staining of sulphate minerals, the X-ray diffractionidentification of deleterious crystalline materials, and infra-red spectroscopy or the use ofscanning electron microscopy plus a microanalyser for the identification of chemicalphases are all very helpful tools in the identification of small amounts of deleteriouschemical components.

The upper limits for soluble chloride specified in BS 882 (1992) range from 0.01 percent to 0.05 per cent depending on the type of concrete. Problems with chloride contaminationof aggregates are rare in the UK since aggregates are usually washed, but in some climaticregions, notably the Middle East, chloride contamination is not uncommon. Chloridereduces the protection of reinforcing steel from corrosion (passivation) and leads to rustforming on the surface of the steel. The rust occupies a much greater volume than theoriginal steel thus disrupting and cracking the concrete as well as weakening the allimportant concrete–steel bond and the steel itself. Carefully measured amounts of gypsum

Table 5.14 Some potentially deleterious constituents found in aggregates

Geology, aggregates and classification 5/27

(hydrated calcium sulphate) are normally added to Portland cement to control its settingcharacteristics, but additional amounts derived from the aggregate could produce deleteriousexpansion and disruption of the concrete through internal sulphate attack. Such damagecan be readily identified in a petrographic thin-section (Figures 5.11–5.13). Solublesulphate is rarely a problem with UK aggregates. However, sulphates such as gypsummay form coatings as is illustrated in Figure 5.14, or may be present as discrete particlesin aggregates and may be a serious problem with aggregate sources in some climaticregions, especially hot dry deserts.

Figure 5.11 A photomicrograph of a petrographic thin-section showing needle-shaped crystals of ettringitein a void, the result of conventional sulphate attack.

Figure 5.12 A photomicrograph showing the effect of sulphate attack which forms the sulphate mineral,thaumasite, which infills the irregular cracks.

5/28 Geology, aggregates and classification

Metal cations such as lead and zinc are known to interfere with the setting of cement.However, problems are rarely encountered in practice and only occur if aggregates havebeen contaminated with soluble metal compounds, or where recycled metaliferous miningwastes have been used as aggregate.

Figure 5.13 A severe example of the thaumasite type of sulphate attack in a hand specimen of concrete.

Figure 5.14 A photomicrograph showing rounded fine-aggregate sand grains with a coating of crystallinegypsum.

Geology, aggregates and classification 5/29

5.7.2 Modification to the strength and durabilitycharacteristics of a concrete

Although small amounts of dust are helpful in concrete, improving cohesiveness andreducing bleeding, excessive amounts may cause the aggregate particles to become coatedwith dust so that the necessary bonding between the particle and the cement matrix is notfully established. As a consequence the concrete will have poor durability and strengthcharacteristics from early ages. Coatings of clay or other weak material can have similareffects.

Interaction between aggregate particles in a concrete or mortar and constituents fromthe cement binder can also lead to longer-term durability problems. The best known ofthese are the alkali-aggregate reactions between certain types of siliceous or carbonateaggregates and the alkalis normally derived from the cement. The alkali silica reactioninvolves aggregates which are composed of (or contain) non-crystalline or poorly crystallinesilica or siliceous glass which can react with the alkali hydroxides derived from thecement. In time, the reaction product which is an expansive hydroscopic gel, can exertsufficient pressure to crack and disrupt the concrete. Certain impure carbonate aggregateswill also react with alkalis in the pore fluid (referred to as the alkali carbonate reaction)and can also cause expansion and disruption of the concrete. Both these reactions betweenalkali and aggregate in concrete are dealt with in Volume 2, Chapter 13 of this series.

The microporosity of aggregate particles may have a significant effect on the strength,durability and freeze–thaw resistance of a concrete. However, it is difficult to generalizesince some porous aggregate materials appear to have beneficial effects when used in aconcrete. The nature of the pore structure and pore size distribution, particularly themicroporosity, appears to be more critical than the overall porosity of the aggregate inrelation to its resistance to freeze–thaw in concrete, so that each aggregate source needsto be assessed individually. Some flints which occur in southern England have a whitemicroporous flint outer layer or cortex which may be present as discrete particles orsurface coatings in an aggregate and these can cause spalling and pop-outs from concretesurfaces under freeze–thaw conditions as is illustrated by Figures 5.15–5.19.

5.7.3 Unsound aggregate particles in concrete

A wide range of aggregate materials contain unsound particles, present either as naturalcomponents in the aggregate source material or as contaminant particles from an extraneoussource. They can produce problems when incorporated into a concrete which range fromsurface cosmetic disfigurement to reduction in the compressive strength, or a reduction indurability.

One of the commonest unsound particle contaminants familiar in the gravel aggregatesof southern Britain is pyrite (iron disulphide). This mineral can oxidize to brown ironhydroxides in normal environmental weathering conditions if close to the surface of aconcrete. The ease with which this oxidation takes place depends on the purity and thedetailed crystal structure of the mineral, but can lead to development of brown staining onthe concrete surfaces illustrated in Figure 5.20. Although such stains do not normallyaffect the structural integrity of dense concrete, the unsightly stain is usually much largerthan the reacting mineral particle.

5/30 Geology, aggregates and classification

However, some Cornish tin mining wastes contain pyrite crystals, or ‘mundic’ particles,(mundic is the old Cornish word for pyrite). This material is used as low-grade aggregatefor the production of concrete blocks and the pyrite does contribute to the deteriorationof concrete, although moisture movement in the clay and the secondary micas which arepresent in the slaty and phyllitic waste is also involved. Examples of the type of deteriorationare shown in Figures 5.21 and 5.22. Guidance notes concerning this particular ‘mundic’problem have been published by the Royal Institution of Chartered Surveyors (1997).

Aggregate particles partly composed of clay and certain geologically weathered rocks

Figure 5.15 Examples of flint and chalk (extreme right) aggregate particles showing the whitemicroporous cortex on the flints (partly broken away on the particle second from left).

Figure 5.16 A scanning electron microscope image of dense flint spares are the black areas.

Geology, aggregates and classification 5/31

tend to be moisture sensitive, that is, they have the capacity to absorb or lose water fromtheir structure depending on the environmental conditions. This produces a consequentexpansion or shrinkage of the particle leading to the breakdown of the aggregate/cementbond so that the concrete becomes susceptible to damage by frost and thermal cycling.This problem of aggregate shrinkage was recognized in Scotland in the 1950s (Edwards,1970). Attempts were made to list the petrological rock types susceptible to shrinkage,but this proved to be too broad and unreliable. The petrological evaluation of individualaggregate materials taking account of factors such as weathering together with testing(for example, BS 812, Part 120, (1990) or ASTM C157-99 (1999)) are necessary toestablish the stability of an aggregate for use in concrete.

Figure 5.17 A scanning electron microscope image of microporous flint.

Figure 5.18 Surface pop-outs caused by microporous flint aggregate in the concrete.

5/32 Geology, aggregates and classification

Figure 5.19 Scaling of a concrete surface owing to freeze–thaw, exacerbated by de-icing salts.

Figure 5.20 Iron staining due to a pyrite particle present near the surface of a concrete.

Geology, aggregates and classification 5/33

Chalk particles present in some gravel aggregate sources (see Figure 5.15) can alsoreduce the quality of a concrete. Chalk aggregate particles, like clay lumps and other softmaterials, are weak and can reduce the compressive strength and durability of the concrete.However, a fraction of the chalk appears to be broken up and becomes involved in thehydration and setting of the cement matrix with uncertain consequences. Thus only aproportion of the original chalk particles remain as true aggregate, nevertheless a smallpercentage of chalk in the aggregate is sufficient to reduce the strength of a concretesignificantly.

Varying amounts of coal and lignite particles are found in some gravel aggregates and

Figure 5.21 An example of a degraded ‘mundic’ block.

Figure 5.22 ‘Mundic’ decay as seen in petrographic thin-section.

5/34 Geology, aggregates and classification

also in some mine wastes. They tend to form weak and porous particles in the aggregate,and may also form unsightly tarry stains on concrete surfaces, or be the cause of surfacepitting. An example of coal aggregate particles in petrographic thin section is shown inFigure 5.23.

Figure 5.23 A petrographic thin-section photomicrograph showing porous coal fragments in a concrete.

The micas, biotite and muscovite, are common flaky constituents of many igneous andmetamorphic rocks. The processes of erosion acting on such rocks tend to produce sandswhich contain varying proportions of discrete mica particles, particularly the colourlessmuscovite mica which is the more stable, and does not readily break down. Mica as aconstituent of a sand fine aggregate usually increases the water demand of a concrete andbecause of flaky shape and surface texture tends to reduce the cohesiveness of the mix.Small percentages of mica in the fine aggregate can reduce the final compressive strengthof a concrete significantly. In one example from south-west England strength was reducedby 5 per cent for 1 per cent by weight of discrete mica particles in the total aggregate.However, such reductions in strength can be offset by using higher cement contents oradmixtures.

Many sea-dredged and some land-based gravel resources contain carbonate shellfragments. Although these fragments are usually mechanically strong complete shellsmay be hollow and other fragments flaky. The nature of these fragments may reduce theworkability of a concrete mix and hollow shells may reduce the strength of the concrete.However, unless the proportion of shell in the aggregate is very high the effects of shellon the workability and strength of concrete are minimal.

References

American Society for Testing and Materials (1988) Standard descriptive nomenclature of constituentsof natural mineral aggregates, ASTM C294–98.

Geology, aggregates and classification 5/35

American Society for Testing and Materials (1999) Testing and Materials (1999) Test for lengthchange of hardened cement, concrete and mortar, ASTM C157–99.

American Society for Testing and Materials (1998) Standard practice for petrographic examinationof aggregates for concrete, ASTM C295–98.

American Society for Testing and Materials (1992) Standard practice for sampling aggregates. (Re-approved 1992), ASTM D75–87.

British Standards Institution (1975) Methods for sampling and testing of mineral aggregates sandsand fillers, BS 812. Parts 1–3.

British Standards Institution (1984) Methods for sampling, BS 812, Part 102.British Standards Institution (1985) Methods for determination of particle size distribution,

BS 812, Part 103.British Standards Institution (1989) Methods for determination of particle shape, BS 812, Part 105.British Standards Institution (1989) Method for testing and classifying drying shrinkage of aggregates

in concrete, BS 812, Part 120.British Standards Institution (1990) Methods for determination of drying shrinkage, BS 812, Part

120.British Standards Institution (1990) Specification for lightweight aggregates for masonry units and

structural concrete, BS 3797.British Standards Institution (1992) Specification for aggregates from natural sources for concrete,

BS 882.

British Standards Institution (1994) Procedure for qualitative and quantitative petrographic examinationof aggregates, BS 812, Part 104.

British Standards Institution (1997) Tests for general properties of aggregates, Part 3: proceduralterminology for simplified petrogaphic description, BS EN 932–3.

British Standards Institution (2002) Aggregates for concrete, BS EN 1260.Carmichael, R.S. (1982) Handbook of Physical Properties of Rocks, Volume 1. CRC Press, Boca

Raton, FLA, Table 17, p. 34.Concrete Society (1999) Alkali silica reaction. Minimising the risk of damage to concrete. Guidance

notes and model specification clauses. Report of a Concrete Society Working Party. ConcreteSociety Technical Report No. 30, 3rd edn.

Douglas, I. (1986) Hot wetlands. In Fookes, P.G. and Vaughn, P.R. (eds), A Handbook of EngineeringGeomorphology, Blackie, Glasgow (Surrey University Press).

Eagar, R.M.C. and Nudds, J.R. (2001) The Geological Column, 8th revised edn. Printguide Ltd,Manchester, Panel 1.

Edwards, A.G. (1970) Shrinkable aggregates. Building Research Station Scottish Laboratory, EastKilbride, SL 1, (1) 1–70.

Folk, R.L. (1959) Practical petrographic classification of limestones. AAPG Bulletin, 43, 1–38.Fookes, P.G. (1989) Civil engineering practice. In Blake, L.S. (ed.), Civil Engineer’s Reference

Book, 4th edn. Chapter 8, p. 8/9.Fookes, P.G. and Higginbottom, I.E. (1975) The classification and description of near-shore carbonate

sediments for engineering purposes. Géotechnique, 25: 406–411.Fookes, P.G., Gordon, D.L. and Higginbottom, I.E. (1975) Glacial landforms, their deposits and

engineering characteristics, in the engineering behaviour of glacial materials. Proceedings of asymposium of Midland Soil Mechanics and Foundation Engineering Society, reprinted byGeoabstracts, Norwich, 1978.

Fookes, P.G. and Higginbottom, I.E. (1980) Some problems of construction aggregates in desertareas, with particular reference to the Arabian Peninsula. I – Occurrence and special characteristics.II – Investigation, production and quality control, Proc. of the Institution of Civil Engineers, Part1, 68, Feb. pp. 39–90.

Fookes, P.G., Dearman, W.R. and Franklin, J.A. (1971) Some engineering aspects of rock weatheringwith field examples from Dartmoor and elsewhere. QJEG, 4, 139–185.

Fox, R.A. (2002) Non-energy mineral resources. In Maritime World 2025, Future Challenges and

5/36 Geology, aggregates and classification

Opportunities, The Greenwich Forum conference proceedings 3–5, April 2002, available onCD-ROM.

Geological Society Engineering Group (1995) The description and classification of weatheredrocks for engineering purposes. QJEG, 28, 207–242.

Kirkaldy, J.F. (1954) General Principles of Geology. Hutchinson, London, pp. 1–327.Leet, L.D., Judson, S. and Kaufman, M.E. (1982) Physical Geology, 6th edn, Prentice Hall,

Englewood Cliffs, N.J, pp. 284–285.Mclean, A.C. and Gribble, C.D. (1990) Geology for Civil Engineers, 2nd, edn. Unwin Hyman,

London.Royal Institution of Chartered Surveyors (1997) The ‘Mundic’ problem – a guidance note –

recommended sampling, examination and classification procedure for suspect concrete buildingmaterials in Cornwall and parts of Devon. 2nd edn (Chairman) Stimson, C.C., pp. 1–75.

Smith, M.R. (ed.) (1999) Stone: building stone, rock fill and armourstone in construction. GeologicalSociety Engineering Geology Special Publication No. 16, London, pp. 1–478.

Smith, M.R. and Collis, L. (eds) (1993) Aggregates, sand, gravel and crushed rock aggregates forconstruction purposes, 2nd edn. Geological Society Special Publication No. 9, London.

Smith, M.R. and Collis, L. (eds.) (2001) Aggregates, sand, gravel and crushed rock aggregates forconstruction purposes, 3rd edn. Geological Society Special Publication No. 9, London.

St John, D.A., Poole, A.B. and Sims, I. (1998) Concrete Petrography. A handbook of investigativetechniques. Arnold, London.

Sims, I. and Brown, B.V. (1998) Concrete aggregates. In Hewlett, P.C. (ed.), Lea’s Chemistry ofCement and Concrete, 4th edn. Arnold, London, Chapter 16.

Winchester, S. (2001) The Map that Changed the World. Viking, Penguin Books, London.

6

Aggregate prospectingand processing

Mark Murrin-Earp

6.1 Aims and objectives

This chapter describes the processes involved in the winning of aggregates and theirprocessing, looking at the different methods employed in doing so and the machinery andequipment used. Predominantly the areas covered will be:

• Extraction and processing of sand and gravel• Extraction and processing of limestone.

It must be realized that generic examples will be used for the layout of differentextractive processes, using typical machinery and equipment and that naturally occurringsources of raw materials vary. Therefore, when approaching a specific deposit, it may benecessary to employ modifications to the types, setup and usage of such equipment. Thesections will be divided into:

• Extraction.• Processing and machinery and equipment.

6.2 Introduction

As long as society demands that buildings and roads be constructed, then naturallyaggregates will be extracted to produce the materials required. The aggregate extraction

6/2 Aggregate prospecting and processing

industry has grown to keep pace with demands for aggregates for use in constructionmaterials and since the Second World War the volume of materials extracted has increasedsteadily, with only brief deferments for periods of global and national recession. Today,modern sophisticated machinery and equipment are used to produce high volumes ofmaterials.

The characteristics and properties of the aggregate can go a long way to dictate thefinal performance of the material. Obviously this will mean differences in the qualities ofaggregate sources and will influence the selection of the aggregate source for its final use.

Aggregate quality, location and mineral type will also determine the mode of extractionand the processing that it will subsequently receive post-extraction.

Aggregate is a mass-produced commodity and therefore must be a cost-effective rawmaterial to the marketplace. However, because mass production techniques are employedto produce cost-effective construction material sources, this does not mean that qualityissues can be overlooked. High levels of capital investment by companies in sophisticatedstatic and mobile machinery and equipment results in high quality and consistent products.

6.3 Extraction and processing of sand and gravel

First, in the case of wet and dry extraction of sand and gravel consideration must be givento the removal of any overburden from the deposit. It is important to realize the value ofaccurate site investigation prior to the commitment to extract.

Other concerns in this area will centre around any costs related to the removal of anyoverburden and the implications of any further sterilization of deposit by the removal andsubsequent stocking of this material, along with its use in any future restoration of thefinished site. This last factor of restoration is again a consideration in terms of the finalcosts of completion of the site.

The extraction of land-won sand and gravel consumes a large amount of land due tothe relatively shallow nature of most deposits of these types. Most land-based depositsare usually only, at worst, semi-bound (relates to the degree of cementation of the deposit)or typically unbound. It is common to strip overburden in a phased manner, in order tominimize the risk of double-handling of material and to remain cost effective. The use ofgood estate management practice is essential if this is to be kept to a minimum.

The three main methods of extracting sand and gravel are as follows:

• Wet• Dry• Marine

Wet extraction refers to a deposit below water table level, with the material to be extractedin a constantly ‘wet state’. This mode of extraction requires equipment that is largelyexclusive to this deposit in the case of suction dredging and floating cranes or grabs.Draglines may be also used, which in turn is a commonly used method to extract ‘drydeposits’.

The use of tracked or walking draglines has certain disadvantages in terms of thedepths of extraction that can be worked, economically with depths of between 5 and 10metres. Loss from the bucket is also a factor. Generally we must consider the amount of

Aggregate prospecting and processing 6/3

additional conveyor line or dump trucks, which may prove problematic when workingsare of any distance from the site allocated from the processing plant.

If choosing standard or multiple-axle articulated dump trucks to deliver won materialfrom a dragline, care must also be exercised so that the correct selection of capacity isachieved. This allows the primary extraction device to have an equal amount of bucketpasses from the face to the dumper so as not cause under- or overloading of the vehicle.This will result in efficient material transfer to the processing plant. Multiple ‘shifts’ ofvehicles may be required to turn around high volumes of won material for processing.Alternatively the material can be temporarily stocked to allow some free draining to takeplace. Generally the use of vehicles to transport material any distance is inefficient and,wherever possible, conveyor transport should be used.

One of the alternatives to the above is using a suction dredger. Sites of this naturewhere this system is to be employed as the means of extraction must be looked at withcare, as they require years of extraction to make the initial investment worthwhile.

Finally the use of floating cranes with a fitted grab or a ladder dredger (inclinedconveyor system which has buckets fitted to allow the material to be scraped from the bedof the submerged deposit) in this environment is also suitable, where a barge-mountedcrane with a grab-type bucket is fitted. Usually to ensure that the operation is effective itis necessary to have some form of surge hopper fitted to the barge, which in turn isconnected to land via a conveyor and pontoon arrangement. This allows the barge tomove around different areas of the submerged deposit easily once an area has beenworked to the desired depth. Again with this type of deposit, a relatively shallow extractiondepth can be achieved. So once again a commitment to a large area of extraction isrequired. Restoration planning on this type of site will usually mean a lake as obviouslyno overburden stripping is required and therefore there will be no surplus materials toeffect a fill. The won aggregate can be pumped ashore or fed into barges, which in turnare then landed to deliver the material for processing.

Generally these types of extraction are on a large scale, with higher volumes of materialinvolved. ‘Who wants to buy a second-hand land-locked boat?’ Anon., 1997).

Dry ‘pit’ extraction refers to deposits that are either naturally dry or which are extractedfrom land that has been drained to facilitate dry extraction. Again in this environmenteither dragline or face shovel extractive means can be employed, depending on the volumerequired. Limitations for draglines again apply, so to be effective, extraction deposits of5 m in thickness or greater are required. These are efficient and are therefore widely used.

Face shovels are another option, and prove efficient in unbound or semi-bound extractiveenvironments.

Variations on hydraulic machinery such as the use of back-acting or backhoe excavatorsalso have advantages and disadvantages within themselves, in that they are capable ofdigging semi-bound dry deposits. The main disadvantage compared to a dragline is thedepth of formation to which they can dig. However they are extremely useful for formingbunds and barriers on-site, as well as assisting in the reinstatement of finished areas.

The size of the pit will depend on the number of hydraulic excavators used. Typicallyin a medium-sized pit producing 100 000 tonnes or more two faces can be worked.Extracted material is then transported to the primary processing plant or, depending onthe site layout, a surge pile or hopper.

Moving the raised materials to processing again tends to favour high-speed conveyors,which are fixed, and they in turn are fed from branch conveyors that allow for movement

6/4 Aggregate prospecting and processing

and re-assembly to the live face working area of the quarry. The most important factor isto minimize the use of vehicular transportation, which in turn reduces operating costssuch as wear and tear on expensive capital items of plant. Again, as with wet extraction,we consider the ratio of the hydraulic loading device to the size of dumper used. Forinstance, if a long-reach back-acting unit is used, in order to reduce the stresses on thearm, a smaller bucket is fitted. Therefore it may require several passes in order to fill alarger dump truck. In this case one needs to carefully design the extraction to maximizethe use of equipment.

When opening a new dry or wet pit, it may also be necessary to use mobile and notstatic plant and equipment in order to begin the works. Such equipment can be installedperiodically to boost production of a particular material and to assist the existing staticplant. This may be done by engaging a suitable contractor, or by hiring the equipmentfrom a plant hire company. These measures are to be viewed as temporary and mayrequire some addressing regarding planning permission to have the site for longer periodsof time.

A wide variety of equipment is available, and if the extraction is for a limited time, thismay well prove an option, particularly if the pit is producing a small range of differenttypes of material in high volume. Considerations in terms of locating this plant on-site arerequired, again as previously stated to maximize output and efficiency.

Marine extraction is by suction dredger and takes place in river estuary environmentsand continental shelves and as a means of extraction this method has grown inpopularity.

Investment is initially high, as construction of a specialist ship is required as well asa suitable wharf for landing of the raised aggregates.

The wharves are also ideally located close to the demand for the materials and withinreach of the source, to effect a collect and discharge cycle. This allows the won aggregateto drain and be processed while the dredger returns to sea to reload. Transport by ship isalso increasing in popularity as the economies of scale come into play, with high volumesbeing transported.

There are two main modes of operation for dredging fleets to win aggregates:

• Dredging at anchor with a forward leading suction pipe• Trail dredging with the ship underway.

Dredging at anchor implies that the ship is anchored and the suction pipe is loweredto the seabed in forward mode. This forms a ‘sea pit’ from the initial lowering andcommecement of the operation. Once this pit has been formed and because the aggregatebeing extracted exhibits fluid properties at this stage, the pit on the seabed is continuallyreplenished from the aggregate immediately surrounding the excavation. Periodically theship will need to move forward to form another pit and the process can then recommence.

The main advantage of this is that it allows for deeper extraction depths to be achieved,but it has environmental advantages and disadvantages as well.

Unlike dredging at anchor, trail dredging involves the ship being underway and, as thename suggests, a trailing suction pipe is lowered to the seabed. As the ship sails, aggregatesare pumped or sucked from the deposit by the use of hydraulic pumps. These pumpscreate a head of vacuum that draws up the aggregate. Depths are somewhat limited withtype of operation, with operating depth between 25 and 45 m.

Aggregate prospecting and processing 6/5

A certain amount of processing, by way of washing of some unwanted fines from theraised material, can take place at sea. Generally at sea processing is not permitted.

Extraction licenses may also be required to remove materials from certain environments,particularly if within close proximity to fishing grounds. However, river estuary dredgingfor navigation purposes can have significant benefits. The main drawback is the potentialquality and overall suitability of such materials.

6.4 Processing

The level of processing affecting sand and gravel deposits will often depend on how thatmaterial was originally won. That aside, there are similar attributes to the eventual processingthat the aggregate will receive.

The purpose of processing aggregates is so they can be used in other products, and thetype and level of processing they receive, ranging from simple screening to screening,crushing and washing, has an influence on the finished quality of the aggregate and itsoverall compliance with relevant standards.

There are three main areas of processing with either of the three means of winningsand and gravel:

• Crushing• Screening• Washing

Crushing will largely depend on the percentage of coarse aggregate in the deposit, i.e.the amount of material deemed to be retained on a 5.0 mm test sieve.

Some deposits, usually dry pit river terrace deposits such as the Thames Valley, haveup to 30%-plus gravel in the deposit. The location of the pit by geological site investigationis important, as knowing the ratio of sand to gravel in the deposit validates the choices tomake regarding the selection of processing plant for the site. These ratios are determinedby digging trial pits on the site before any quarrying takes place. Subsequent samplesfrom these trial pits are graded in the laboratory so that an overall estimation can be maderegarding the levels of coarse and fine aggregates in the deposit. If there were an over-abundance of coarse aggregate in the deposit, for instance, this would probably indicatethe unsuitability of extraction taking place from this location.

Also the level of processing of the materials raised from the deposit can depend on thedegree of the bound state of the aggregates.

The most effective method of moving aggregates around the site is undoubtedly bymechanized conveyors. High-speed conveyors can move large volumes of materials inshort spaces of time, and the selection of type and size of the system will depend on itslocation within the process. Material from the primary crushing unit will therefore obviouslyrequire a wider, high-volume conveyor than, say, one moving material in between washingor screening operations. A conveyor is essentially a straight, continuous section of vulcanizedrubber layers bound onto nylon webbing and is extremely durable. This belt is tensionedand held in place by the head rollers and kept lifted by smaller support rollers. Adjustmentof the head roller tension causes the belt on the roller to skew in different directions, sothis makes it possible to keep the unit well maintained and running true. Failure to ensurecorrect belt alignment will cause premature wearing of the belt and additional maintenance

6/6 Aggregate prospecting and processing

costs. If moving materials on an incline the belt can have V slots vulcanized to the beltto allow the material being moved to grip the belt. Head and tail main drums can also befitted with a scraping device to prevent the build-up of detritus on the belt, in the sameway that head and tail rollers are required to be kept clean to avoid uneven wear and tearon the unit.

Aggregate is moved in different directions by the use of transfer points, which areeffectively smaller hopper-like enclosures that prevent material spillage around the point.Under health and safety regulations there will need to be guarding of the moving partsand a means of tripping the conveyor in the event of an emergency stop being required.

Crusher selection and types are, again, largely dependent on the ratio of coarse andfine aggregates present in the deposit and the distribution of particle sizes within it.

It is also possible to carry out some pre-screening of the as-raised aggregates in orderto remove undesired elements such as large cobbles, wood, lignite and clay nodules. Thisis specialist processing that requires particular specific equipment relative to the contaminantin the deposit and will be covered by a later section.

Once the primary considerations for the type of deposit in question have been consideredthen the type of crusher for the reduction in size of the coarse aggregates in the depositcan be selected.

There are three main types of crusher to select from, with some variations on theirindividual themes:

• Jaw crusher• Impact crusher• Cone crusher

A jaw crusher functions by having a fixed jaw and a moving jaw, the material exitingthe crusher at the desired size by setting the gap between the two jaws. The exit gap is setby means of adjusting a bolt and spring assembly at the unit’s base, hence allowingrelatively fine adjustment to the exit gap and the size of the material exiting. This isknown as the reduction ratio.

The crushing motion is by means of a flywheel that causes the moving jaw to swing.The moving jaw can have either single or double toggles fitted which give it additionalmovement in either two or three planes. This gives the moving jaw greater flexibility andis less prone to jamming with oversized feed. The crushing motion is that of a powerfulcompressive force caused by the reciprocal motion of an eccentric shaft.

The jaw crusher is almost a universal design and is found and indeed is suitable inmany locations. It is often used as the primary means of crushing in many pits andquarries and is best choke-fed. Careful selection of the primary feed material for this typeof crusher is vital, as they are prone to jamming with oversize. It may be necessary to usea pre-crushing screen to avoid this. Some crushers of this type can be fitted with pneumaticbreakers, in order to unblock any jams that occur during the production cycle. The wearparts are usually of highly wear resistant manganese steel.

The main advantage of this type of crusher is that they are hard-wearing (requiringbreaker plates on the jaws to be periodically replaced: this will depend on the abrasivenature of the material being crushed, and the general rule of thumb on this is the lowerthe aggregate abrasion value, the more wear and tear will be on the crusher’s wear parts).They are easy to maintain and are a cost-effective medium volume solution to aggregateproduction.

Aggregate prospecting and processing 6/7

The disadvantages are that they cannot achieve very high volumes, they produceaverage quality shaped aggregate, which tends to be elongated, and they are prone tojamming, which causes downtime in the production day.

They can be fed directly from the primary separation of coarse and fine aggregatefractions screening process.

With regard to maintenance, it is usually prudent to carry a spare set of jaws in storeas part of any successful extraction business’s planned preventative maintenance scheme.It is also possible to effect temporary repair to the wear parts of the jaws by applyingstrips of weld vertically to the wear part plates, but only as a temporary measure. Thesetypes of crusher are also relatively cost-effective to run in terms of power demand oncethey are up to speed. The greatest power demand is during start-up, when the motors willhave their highest current draw. The flywheel has an eccentric drive from the standard V-belts that come from the motor; the amount of belts depends upon the size of the crusher.

Sometimes called ‘cone’ crushers, gyratory crushers give a medium output in terms ofprocessing and, unlike jaw crushers, provide an excellent means of shaping aggregate dueto their action. They consist of a lined external cone and an external cone set on aneccentric shaft bearing to produce the crushing motion. However, the shaft does not rotatebut gyrates instead. This makes good compatability for reduction in size of hard andabrasive rock types. Only certain ratios of reduction can be applied, with the reduction oflarger to smaller sizes (40 mm and 28 mm to 14 mm and 10 mm) an ideal choice.

They are best located as part of a secondary crushing arrangement, with a means ofdiverting other sizes via a chute mechanism to allow diversion from the screening process.They are rarely used in primary mode like other types of crusher due to the relatively lowvolumes produced; the entrance aperture at the top of the crusher is also the limitingfactor as to the size that can be fed.

In terms of maintenance, the cone and lined wear parts consist of manganese steel. Ifnot set to gyrate at an optimum amount of cycles (usually between 100 and 200 revolutionsper minute) and feeds are not correctly balanced, this will cause uneven wear andpoor performance. Once set to run correctly they can run for long periods withoutadjustment.

Adjustment of the exit gap for the finished aggregate to be fed to the screening processcan be via either mechanical or, with more modern units, hydraulic. This involves closingdown the exit aperture by either of the above means and allowing the aggregate to beretained longer in the crushing chamber. The more accurate way to control the exit shapeof material is, again like the jaw crusher, to steadily choke feed it. Overfeeding will meanspilt and lost material, while underfeeding will cause low output and poor shape. Ingeneral, all new gyratory crushers have hydraulic gap-setting devices, which is again lesslabour-intensive.

Impact crushers can operate in two planes, horizontally and vertically. They are high-volume devices, often found located as the primary crushing unit. The crushing action isby a rotor, with ‘swing-out’ blow-bars fitted. When fed they ‘throw’ the primary feedmaterial against a series of breaker plates. The exit gaps, which can be a series of bars,prevent the material exiting before it has reached the desired size. An impact crusher is,like the gyratory crusher, conducive to producing good shaped almost cubical aggregates.

The breaker plates, speed of the rotor and blow-bar type are all adjustable, whichmakes this a versatile device. Adjustment of the breaker plates is by hydraulics, whichmove the plates further or nearer the rotor. The nearer the plates are to the rotor, the more

6/8 Aggregate prospecting and processing

energy is dissipated into the crushing action. This will mean a smaller product if desiredor, conversely, a longer time to produce, with longer spent in the crushing chamber.

Horizontal impact crushers have one main disadvantage in that they are prone to wear.Changing the breaker plates can either be by turning round the plates, or when they havebeen used, by new plates. Because there are so many plates employed, this takes timeand, ultimately, will mean total replacement of the rotor itself due to excessive wear andtear of extremely abrasive rocks. However, in moderately abrasive environments, theydemonstrate good wear and value for producing volume aggregate from the primary feedto the screening process. For extremely high volumes of throughput, horizontal impactcrushers can have double rotors fitted.

As the name implies, vertical shaft impact crushers are fed vertically but the crushingaction is this time by throwing the aggregate feed in a horizontal plane. Aggregate isallowed to build up on the sides of the box, and an aggregate wall is built. This now meansthat aggregate is thrown against aggregate and little contact between metal wear partsurfaces occurs. This is a particularly cost-effective means or providing a high-volumefeed to the screens.

They are suitable for crushing higher yields of sand in flint gravels and when crushinggrit-stones produce a good shape. Stone against stone crushing ratios can be higher forthis type of device.

The screening procedure allows the production of aggregate products for componentuse in other products. There are many different methods of screening and, like the crushingoperation can to a certain degree be tailored to suit a particular type of material.

Screening forms part of the secondary process and can involve the product, in particularthe finer aggregate, being washed to allow its use if from a deposit that requires this.

As previously mentioned, some screening of the primary extracted aggregate can takeplace to allow for the removal of any undesired elements or to remove oversized material.This is done by the use of a set of inclined parallel bars that are vibrated; the gap widthof these bars can be varied to allow for different sizes to be removed at source. Thisprocess, or type of screen, is commonly called a ‘grizzly or bar’ screen.

Consideration of choice of equipment must be given; this will depend on the types ofaggregate sizes required and their volumes present in the deposit. For instance, if only thebasic elemental aggregates of concrete production are required, then the amount and typeof plant should reflect this.

Some deposits may include partially fossilized wood and lignite (low-grade coal).Specialist, dedicated plant, by means of density separation, can remove these. In the caseof these contaminants (which have a detrimental effect on any finished concrete) is by theuse of water.

The use of varying sized mesh apertures, which can be rounded, rectangular or square,depending on their desired function, are utilized to produce a finished aggregate productand remove any undersize or oversize from such products. This is done by attaching thescreens to an inclined oscillating, vibrating device, known as the screen deck. This is atypical screen arrangement, where the angle of inclination is between 15–20° and issometimes known as a ‘Niagara’ system. This action creates a backflow effect and allowsthe material to present to the screen, over which it is travelling, many times during theprocess. There are other methods of suspending screen meshes of various types, such asthe arrangement mentioned above, known as ‘multiple layer’ decks. These greatly improvethe area available and capacity of the machine. Other methods include a curved multiple-

Aggregate prospecting and processing 6/9

deck arrangement, as mentioned for pre-screening, to remove oversize, a set of parallelbars and, in the early days of relatively low demand and production levels on the industrya complete round barrel, fitted with curved screens. This process was excellent for producinggood shaped aggregates and was known as a Trommel screen. Due to high productionlevels, this type of screen is rare in the modern production unit.

The efficiency of screening can depend on several factors. The rate of feed to thescreen, the weather and state of wetness of the aggregate all play a part in the sizingprocess as well as the choice of apertures for the screen decks. Also playing a role is thechoice of material from which the screen is manufactured. The original particle shapealso determines the process efficiency, and this in turn can also have an influence on thetype of screen deck chosen. The more rounded the particle, the more difficult the screeningprocess may be in terms of how many particles pass the required aperture.

Screen decks can be made from woven wire (as in test sieves used within a laboratory)or from durable plastics. Wear and tear on the screen decks also depends on the abrasivenature of the aggregate being extracted. The use of woven wire decks in very wet processingmay also fail due to rusting of the decks. Regular checks on the screens should be carriedout in order to avoid total screen failure and the subsequent aggregate produced notmeeting the required specification. They are relatively easy to maintain and replace onceworn and are usually slotted or bolted into place on the main deck of the screening device.This will cover the coarse aggregate fraction of the material extracted from the deposit.The fine aggregate itself undergoes different procedures to achieve the required finish.

There are two main terms that are associated with screening efficiency: pegging andblinding. Pegging refers to the coarse aggregate particles becoming stuck in the screendeck matrix apertures. Excessive pegging will result in lost efficiency and will result inless area available to process the feed and in poorly graded produced aggregate. The useof tensioned wire decks here is an option to overcome this as the vibration of the screendeck over a large area can have some self-cleaning properties which can avoid thisproblem to a certain extent. Large decks with this arrangement, however, once failed dueto wear, have a catastrophic effect on the finished aggregate. Blinding is usually a directresult of higher than normal levels of moisture in the aggregates being screened and thesubsequent agglomerations that build up on the screen mesh. In turn, this causes a reductionin the area available for the process and an overall reduction in capacity. Inclined screensare particularly prone to this if the amplitude of vibration is insufficient and when themoisture content reaches a given level the blinding effect will take place. The use ofvarying size pulley wheels can be used to adjust the turn rate of the off-set cam thatproduces the vibration. Both the pegging and blinding effects can be avoided by regularchecks on the quarry screen house and to a degree prevented by good maintenance.

The next part of the process in terms of the fine aggregate from the deposit is classification.This is done by turning the material into a fluid state by the addition of water. Injectingwater at pressure causes a flow current and allows different sized particles to be extractedat certain levels within the classifier. The particles are then separated or ‘classified’according to their mass, so the upward current produced by the classification processmeans that the fine sand will be taken from the top and, due to its greater mass, the coarsesand at the bottom.

At this stage of the process the classified sands are very wet and therefore need toundergo a process known as de-watering. This can either be by simply allowing thematerial to be stocked out and to free drain, due to a process of stock management or by

6/10 Aggregate prospecting and processing

means of a dedicated de-watering plant fitted after classification. In general, the finer thesand particles, the greater the amount of time that is required for the material to drain.This is due to the increased pore pressure that builds up within the finer sands and a de-watering plant breaks these pressures by using spaced strips of a material with elastomericproperties, which is fixed together with horizontal tie bars. The assembly is then vibratedand a thick layer of the sand is allowed to build up over the inclined screen deck. Thesubsequent vibrations of the screen cause the pore pressures and surface tension effect tobe broken and the water allowed to pass through the material.

There are some variations to the method of classification; mainly the use of hydrocyclonesand Archimedian screw type devices. The hydrocyclones are effectively a typical classifierarrangement and the screw types involve the movement of material up an inclined poolfrom the point where the sand feed enters. Again this is a type of density separation, themain disadvantage of this type of device being that generally they do not suit largerproduction volumes.

The amount of water removed is important for the eventual use of the product as a rawmaterial, as excessive moisture contents can lead to difficult product quality control inend use.

6.5 Extraction and processing of limestone

The use of limestone in the production of concrete will depend on the location of thesource, the availability of other sources and the type of concrete to be manufactured withthe aggregate. The use of explosives to extract the primary material for processing is thefundamental difference between the extraction of sand and gravel to what are commonlycalled ‘hard rocks’ by the industry. This therefore adds an extractive additional cost to theproduct (due to the additional use of explosives training, supervision, storage and use).These types of aggregates are also planned for on a long-term basis and are land-wononly. This factor itself attracts different levels of planning permission to extract and theprocess of restoration of the finished works will also require some degree of specialisttreatment.

The material is extracted by explosives by using a drilling rig to make a series orpattern of holes on the top of the quarry face. Each face (and there are usually several thatmake up a quarry) is known as a bench and allows the material to be extracted in stagesby progressively moving down a stage. An individual quarry may have several faces witheach bench in a face being from 3 to 5 metres in height. It is important to divert as muchof the energy from the explosives into the surrounding rock to cause a high degree ofshatter and to avoid large unusable ‘pop’ rock that cannot be fed to the primary crushingdevice because of its size.

A drilling rig with a percussive action is used to drill a series of hole patterns withinthe quarry face on the bench to be blasted. The holes follow the profile of the face and arelaid down in echelon. These are then charged with explosives, which are largely purchasedfrom suppliers rather than manufactured on-site, called ANFO (ammonium nitrate ferrousoxide), a mix of gas oil and industrial grades of fertilizer.

If there is any part of the face which has been deemed to have softer, less resistantrocks, then these areas will not require charging with explosive and must be stemmed.This means that small aggregates are poured into the hole to avoid an escape of energy

Aggregate prospecting and processing 6/11

into the less resistant area and thus causing the potential risk of rock flying away from theface and resulting in damage or injury. It has been recorded that a piece of rock of some200 mm in diameter has travelled up to 2 km from the quarry face. A method of furtherreducing this risk is to cover the face to be blasted with disposable netting that will catchany escaped rock.

The holes are also run down with the detonators, which are linked to the rest of thehole in the ‘shot’ to be ‘fired’. Each detonator or ‘cap’, as they are sometimes known, hasa specific detonation time and the plan is usually to time the operation so that the frontof the blast (i.e. the part nearest the face) is fired first, followed by the rear. This takesplace in milliseconds from the shot being fired.

Each blast can bring down several thousand tons of primary rock at a time and howmany times a quarry blasts will depend on the volumes required to fulfil the needs. Twiceto three times a week is not uncommon in an average sized quarry producing up to amillion tons per year.

Any large unsuitable rock from the blast can be separated and later can be reduced bythe use of a hydraulic pick fitted to a back-acting loader. (These are similar to the breakerthat can be fitted to a primary crusher usually of the jaw variety: alternatively they can besold for use as coastal or sea defence works.) Secondary, tertiary and screening processesutilize the equipment as described above (see typical layouts for these types of extraction).

6.6 Summary

The extractive minerals industry has developed rapidly from the end of the Second WorldWar to accommodate the growth of society and the need for an improved infrastructurein a world economy. Each stage of the processes and the equipment required to extractand process raw materials into finished usable aggregates has been examined, and varyinglocations will require the combination of several of these processes to achieve this. Theoperation of an extraction site requires knowledge not only of the equipment and processesinvolved but also of the raw material itself, planning issues, health and safety and theenvironment. Other aspects not covered by this chapter are those of engineering andelectrical installation and operation of the associated equipment with the extractive industry,and the author recommends a broad knowledge of these matters.

The impact of extraction of raw materials on the environment also has to be consideredalong with the responsible operation of sites, their use by future generations and anyfuture recycling of other sources.

Further reading

Anon. (1999a) New Parker plant for Hereford quarry. Quarry Management. December.Anon. (1999b) Mobile Plant. Quarry Management. August.British Standards Institution, BS 63 Part 1, 1989.British Standards Institution, BS 63 Part 2, 1989.Mellor S.H. (1990) An Introduction to Crushing and Screening. Institution of Quarrying, Nottingham.QPA. (1995) Statistical Year Book Quarry Products Association, London.Smith, M.R. and Collis, L. Aggregates (2nd edn). The Geological Society, London.

7

Lightweight aggregatemanufacture

P.L. Owens and J.B. Newman

7.1 Introduction, definitions and limitations

Lightweight-concrete is defined by BS EN 206-1 as having an oven-dry density of notless than 800 kg/m3 and not more than 2000 kg/m3 by replacing dense natural aggregateseither wholly or partially with lightweight aggregates. Although excluded from this Standardthis range of densities can be achieved if the concrete is made with most dense naturalaggregates but in such a way that excess air is incorporated as, for example, in no-fines,aerated or foamed concrete.

Lightweight aggregate is defined in BS EN 13055 as any aggregate with a particledensity of less than 2.0 Mg/m3 or a dry loose bulk density of less than 1200 kg/m3. Theseproperties mainly derive from encapsulated pores within the structure of the particles andsurface vesicles. For structural concrete suitable aggregates should require a low cementpaste content and have low water absorption.

The most appropriate method of assessing particle density in structural concrete is tocompare the density of the particle to that of typical natural aggregate, e.g. with a ‘reference’particle density of 2.6 Mg/m3. The minerals comprising the solid structure of most aggregates,whether lightweight or not, have densities close to 2.6 Mg/m3, but it is the air within thestructure of an aggregate that enables compacted structural concrete to be less than2000 kg/m3.

It is important to understand the difference between the two definitions of density.Particle density relates to the mass per unit volume of individual aggregate particles asmeasured by BS EN 1097-33. Dry loose bulk density is the mass of dry particles contained

7/2 Lightweight aggregate manufacture

within a given volume as measured by BS EN 1097-64. To demonstrate this differenceFigure 7.1 illustrates the relationship between these two definitions of density in relationto particle shape, texture and interstitial voids. It is thus apparent that those aggregateswith the lowest interstitial void space are those which give the lowest concrete densities.

Figure 7.1 Relationships between dry loose bulk density and particle density for manufactured aggregate.

30

40

50

60

70

Inte

rstit

ial v

oids

(%

)

Spherical/s

mooth

Irregular

Particle shapeand texture

Angular/vesicular

1400

1200

1000

800

600

400

200

0

Dry

loos

e bu

lk d

ensi

ty (

kg/m

3 )

0.5 1.0 1.5 2.0Particle density (Mg /m3)

7.2 Lightweight aggregates suitable for use instructural concrete

For structural concrete, the requirements for lightweight aggregate are that it:

• has a strength sufficient for a reasonable crushing resistance (BS EN 13055)• enables concrete strengths in excess of 20 MPa to be developed, and• produces compacted concrete with an oven-dry density in the range 1500–2000 kg/m3.

For concrete using fine aggregate from natural sources it is the particle density of thecoarse aggregate that has the greatest influence on the oven-dry density of compactedconcrete. To achieve lightweight concrete the particle density of the coarse aggregateshould be less than about 0.65 Mg/m3 at the lower end of the density range or greaterthan about 1.85 Mg/m3 at the higher end. Typically, an aggregate particle density of1.3 Mg/m3 would produce concrete with a fresh density of about 1750 kg/m3.

Traditionally, manufacturers of lightweight aggregates have been constrained by notonly the limitations of the raw material and the method of processing but also the requirementsof the market. The properties of lightweight concrete, and hence lightweight aggregate,can be exploited in a number of ways from its use as a primarily structural material to itsincorporation into structures for the enhancement of thermal insulation. This variety ofpurpose is recognized by RILEM/CEB who proposed the classification shown in Table 7.1.

This chapter relates to the manufacture of aggregates for concretes within Class I (i.e.structural lightweight concrete).

Lightweight aggregate manufacture 7/3

For aggregate manufacture, the availability of the mineral resource will be crucialsince the lower the finished aggregate particle density, the less will be the rate at whichthe resource will be depleted for the same volume of production. This creates an obviousconflict of interest – does a producer develop a lightweight aggregate for a market whichdepletes the resource at a greater rate or for a speculative market with the possibility ofa smaller return on capital? This dilemma has caused more lightweight aggregate producersto fail than in any other segment of the aggregate industry.

7.3 Brief history of lightweight aggregate production

The history of lightweight aggregate production from natural sources dates back to pre-Roman times and continues today with porous rocks of volcanic origin. However, thesources are limited to regions of volcanic activity.

From around the end of the nineteenth century, with the development of reinforcedconcrete and the rarity of natural porous aggregate deposits and their non-existence inmost developed countries, research for the manufacture of aggregates commenced. InEurope, in the early part of the twentieth century development concentrated on foamingblastfurnace slags since steel production was basic to the industrial infrastructure. However,it was not until the early 1970s that significant developments in pelletizing and expandingblastfurnace slags took place, so that today a slag-based aggregate with a smoother non-vesicular surface, more adaptable for structural concrete, is produced.

It was not until about 1913 that research in the USA revealed that certain clays andshales expanded when fired. In about 1917 this led to the production in Kansas City,Missouri in a rotary kiln of a patented expanded aggregate known as ‘Haydite’ which wasused in the construction of the USS Selma, an ocean-going ship launched in 1919. Therefollowed in the USA the development of a series of aggregates known as ‘Gravelite’,‘Terlite’, ‘Rocklite’, etc. In Europe, however, it was not until 1931 that the manufactureof LECA (lightweight expanded clay aggregate) commenced in Denmark. Thereafter,developments rapidly spread to Germany, Holland and the UK Rudnai (1963). The principleof expanding suitable argillaceous materials, such as the geologically older forms of clay(e.g. shale and slate), has been exploited in the UK since the 1950s with the productsbeing marketed as ‘Aglite’, ‘Brag’, ‘Russlite’ and ‘Solite’. All these companies failed forone reason or another, mainly due to the inaccessibility of the market, non-homogeneityof the feedstock, undesirable gaseous and particulate emissions and the high cost ofproduction, but never on the technical performance of the aggregate in concrete.

In the 1950s the UK Building Research Establishment (BRE) developed the technologyfor the production of a high-quality lightweight aggregate based on pelletized pulverized-fuel ash (PFA) resulting from the burning, generally in power stations, of pulverized

Table 7.1 RILEM/CEB classification of concretes (RILEM, 1978)

Class I II IIIType Structural Structural/Insulating Insulating

Compressive strength (N/mm2) >15.0 >3.5 >0.5Coeff. of thermal conductivity (W/m.°K) n/a <0.75 <0.30Approx. density range (kg/m3) 1600–2000 <1600 <<1450

7/4 Lightweight aggregate manufacture

bituminous or hard coals (Cripwell, 1992). PFA is the residue of contaminants in hardcoals that are present as the result of erosion of natural minerals which sedimented intothe coal measures as they formed. Thus, there is a connection between PFA and otherargillaceous minerals, except that most of the PFA has already been subjected to temperaturesin excess of 1250°C, which vitrifies and bloats some of the particles which are known ascenospheres. Two construction companies became involved in the exploitation of PFAand the knowledge obtained by BRE. Cementation Ltd operated at Battersea PowerStation with two shaft kilns which failed and John Laing & Co. Ltd at Northfleet with asinter strand which became the most successful method in the UK of producing structuralgrades of lightweight aggregate from PFA. It was, and is, marketed under the trade nameof ‘Lytag’ and has an ability to reduce the fresh density of concrete to about 1750 kg/m3

when using Lytag fines and about 1950 kg/m3 with natural fines. These fresh densities areequivalent to oven-dry densities of about 1575 kg/m3 and 1825 kg/m3 respectively.

For the production of structural concrete with an oven-dry density of lessthan 1800 kg/m3, a German manufacturer has produced from about the early 1970s aspherical expanded shale aggregate (‘Liapor’) in a range of particle densities from 0.80to 1.70 Mg/m3. This represents one of the most significant advances in lightweightaggregate manufacture since the aggregate particle can be designed to suit a range ofoven-dry concrete density requirements, thereby giving greater versatility of application.

Other developments are taking place in lightweight aggregate manufacture, such asthe production of hybrids using PFA or other pulverized materials and suitable argillaceousminerals (clay, shales and slate). There are also aggregates termed ‘cold bonded’ whichare mixtures of PFA with lime or Portland cement. Although the manufacture of theseaggregates is now becoming more successful, their use is more appropriate to the productionof masonry than to structural concrete.

7.4 Manufacturing considerations for structural gradesof lightweight aggregate

7.4.1 Investment

For any lightweight aggregate, considerable investment in manufacturing plant is required,sustainable quantities of appropriate resource material must be available and there mustbe a market. In the USA, most cities are based on the principle of high-rise developmentwhich in many cases leads to the use of lightweight aggregate concrete. Europe, with itsmore historic traditions and older infrastructure, has lagged behind. However, moreconsideration is being given to the conservation of land-based mineral resources and it ismore difficult to obtain permission for mineral extraction. The introduction of taxes onnatural materials, which will increase, and the ever-increasing cost of transport, etc. willprovide a further impetus for investment in aggregate manufacture.

7.4.2 Resource materials

The most important asset for any lightweight aggregate manufacturer is a sustainablesource of raw material in a form and state ready for immediate use. Manufacturers using

Lightweight aggregate manufacture 7/5

either PFA or molten slag have immediate advantages compared with those using otherresource minerals. However, the limitations with these materials are that the process ofsintering fuses the PFA particles in such a way that it densifies the agglomerate, whilethe difficulty of entraining air into molten slag limits the particle density to about1.75 Mg/m3.

Alternatively, an aggregate based on argillaceous materials such as clay, shale or slatecan have its density varied by the manufacturing technique. The lighter the aggregate, thelower the demand for resources whereas for the stronger and denser aggregates theresource is depleted at a greater rate. A solution is to make a hybrid aggregate based onPFA mixed with between 25 to 75 per cent by mass of clay (British Patent SpecificationNo. 2218412) which can be heated or fired to make an aggregate with a particle densityof, say, 1.35 ± 0.5 Mg/m3

7.4.3 Processes of lightweight aggregate manufacture

Most manufacturing processes for lightweight aggregates, with the exception of processesusing blastfurnace slag, have been limited to the use of either a sinter strand or a rotarykiln. In instances where the aggregate particle is formed as a friable fresh pellet beforefiring the sinter strand is preferred. Where the form of the fresh pellet is cohesive and itsshape can be retained, the rotary kiln produces the most spherical particle with the mostimpermeable surface.

7.4.4 Production techniques

The various production techniques rely either on agglomeration or expansion (bloating).Agglomeration takes place when some of the materials melt at temperatures in excess ofabout 1100°C and the particles that make up the finished aggregate are bonded togetherby fusion. Alternatively, expansion develops when either steam is generated, as in thecase of molten slag, or a suitable mineral (clay, shale or slate) is heated to fusion temperature,at which point pyroplasticity occurs simultaneously with the formation of gas whichbloats the aggregate.

When appropriate argillaceous materials are heated by firing to achieve appropriateexpansion, the resource mineral should contain sufficient gas-producing constituents andreach pyroplasticity at the point of incipient gas formation. Gas can be developed by anumber of different reactions, either singularly or in combination, from the following:

(a) volatilization of sulphides from about 400°C(b) decomposition of the water of crystallization from clay minerals at approximately

600°C(c) combustion of carbon-based compounds from approximately 700°C(d) decarbonation of carbonates from approximately 850°C(e) reaction of Fe2O3, causing the liberation of oxygen from about 1100°C.

Most argillaceous materials that are suitable become pyroplastic at between 1100°Cand 1300°C. However, depending on the actual source of the material and its chemicalcomposition, the temperature at which bloating for each material becomes effective is

7/6 Lightweight aggregate manufacture

within a relatively small range, usually about ±15°C. At this point the bloated materialhas to be removed immediately from the firing zone and cooled quickly to ‘freeze’ theparticle at the desired degree of bloat, otherwise it will continue to expand. Ultimately, ifexpansion is uncontrolled, the pore wall becomes too thin and there is insufficient resistanceto crushing and the particle will not be strong enough to resist the stresses achievedwithin structural concrete.

A principle of success with all lightweight aggregate manufacture is homogeneity ofthe raw material source, as variability inevitably causes fluctuations in manufacture andthe finished product. To emphasize this present-day aggregate manufacturers have goneto considerable lengths to ensure homogeneity of the raw material.

7.5 Production methods used for various lightweightaggregates

7.5.1 Foamed slag aggregate

In this process molten blastfurnace slag at more than 1350°C is usually poured onto afoaming bed consisting of a large number of water jets set in a concrete base. The waterimmediately converts to steam on contact with the molten slag and penetrates into thebody of the material, at which point the steam becomes superheated. Owing to the rapidexpansion that then takes place, the slag foams to form a cellular structure and vesicularsurface texture. Alternative methods of expansion include spraying water onto the moltenmaterial when it is being tapped from the blastfurnace so that the material is cooledrapidly, with steam becoming entrapped within the structure of the particle. At the completionof foaming the slag is removed and stockpiled, from where it is subsequently crushed andgraded to size. The aggregate produced is very angular with an open vesicular surfacetexture (Figure 7.2).

Figure 7.2 Typical foamed slag particles.

7.5.2 Pelletized expanded blastfurnace slag aggregate

The process of slag pelletization was developed in the early 1970s in order to overcomeenvironmental concerns associated with the production of foamed blastfurnace slag on

Lightweight aggregate manufacture 7/7

open foaming beds or pits. Not only does the slag pelletization process overcome theseconcerns it also produces an aggregate with a closed surface.

To manufacture this aggregate, molten blastfurnace slag at a temperature of about1400°C passes through a refractory orifice or ‘block’ to control the rate of flow. It thenflows onto an inclined vibrating plate with water running down its surface. The vibrationof the plate breaks up the slag flow and traps water which immediately vaporizes andexpands the slag.

More water is sprayed onto the surface of the slag at the end of the vibrating plate. Thisenables gas bubbles to form in the body of the slag, creating further expansion while alsochilling its surface. After being discharged from the vibrating plate the expanded globulesof semi-molten slag are fed onto a horizontal rotating drum fitted with fins which projectthe material through a water mist. The trajectory of material is such that the slag formsirregular and more spherical pellets which are sufficiently chilled to avoid agglomerationwhen they come to rest. After pelletization the material is removed, allowed to drain and,finally, screened to the required size.

The process produces a finished product that comprises semi-rounded pellets with asmooth surface encasing a glassy matrix and a discrete cellular structure, which is essentiallynon-absorbent (Figure 7.3)

Figure 7.3 Typical ‘Pellite’ particles.

7.5.3 Sintered pulverized fuel ash (PFA) aggregate

This aggregate is produced from PFA which is a powder by-product of pulverized bituminouscoal used to fire the furnaces of power stations. Suitable PFA of not more than about 8–10 per cent loss on ignition which results from unburnt carbon in the form of coke (char),is first homogenized in bulk in its powder form. Once homogenized, it is then conditionedthrough a continuous mixer with about 12–15 per cent of water and, as necessary, anamount of fine coal is added to make up the fuel content to about 10 per cent of the drymass of the pellet to enable it to be fired. This conditioned mixture of PFA is then fed ata controlled rate onto inclined and revolving pelletizing dishes. The size and degree ofcompaction of the formed ‘green’ pellets depend on the inclination and speed of rotationof the dish, the rate of addition of the conditioned PFA as well as a further amount ofwater spray. The formed pellets discharge from the pelletizer at a diameter of about 12–14 mm. Without any further treatment these pellets are conveyed to the sinter strand

7/8 Lightweight aggregate manufacture

Figure 7.4 Typical sintered PFA particles.

where they are fed by spreading to form an open-textured and permeable bed to the widthand depth of the sinter strand. The sinter strand is a continuously moving conveyorcomprising a series of segmented and jointed grates through which combustion air can bedrawn to fire the pellets and combustion gases exhausted. Once on the bed, the sinterstrand immediately carries the pellets under the ignition hood that fires the intermixed‘fuel’. The chemical composition of PFA resembles that of clay but, unlike clay, as thePFA has already been fired, no pre-drying or pre-heating of the pellets is necessary, as thepellet is able to expire the water as vapour and combustion gases without incurringdamage to the particles. Once ignited at about 1100°C, and as the bed moves forward, airfor combustion is drawn by suction fans beneath the grate. The process is controlled toprevent the particles of PFA becoming fully molten so that (a) they coagulate sufficientlyto form an aggregate and (b) the aggregate particles are only lightly bonded to each other.The correct amount of coagulation within the pellets is obtained by varying both thespeed of the sinter strand and the amount of air drawn through the bed of pellets.

The finished product on the sinter strand is a block of hard brick-like spherical nodules,lightly bonded by fusion at their points of contact. As the sinter strand reaches the end ofits travel and commences its return to the feeding station a segment of the bed of thefinished product is discharged into a breaker. This action separates the aggregate pelletsprior to grading.

While the surface and internal structure of the finished pellet (Figure 7.4) is essentially‘closed’ it contains encapsulated interstices between the coagulated PFA particles. Whilethese interstices are minute they are penetrable by about 20 per cent moisture but eventually‘breathe’ sufficiently to allow any water to evaporate even when encased in concrete.

7.5.4 Expanded clay aggregate

Expanded clay aggregates are manufactured in rotary kilns each consisting of a long,large-diameter steel cylinder inclined at an angle of about 5° to the horizontal. The kilnis lined internally in the firing zone with refractory bricks which, as the kiln rotates,become heated to the required temperature and ‘roast’ the clay pellets to achieve therequired degree of expansion. The length and configuration of the kiln depends in part onthe composition of the clay and length of time it takes to ‘condition’ the clay pellets in thepre-heater to reach a temperature of about 650°C to avoid them shattering before becomingpyroplastic.

The clay is dug and, to eliminate natural variability, is usually partly homogenized bylayering into a covered stockpile with a spreader before it is removed from the stockpile

Lightweight aggregate manufacture 7/9

by scalping with a bucket conveyor. In the process, a high degree of blending is achieved.The clay is prepared by mixing thoroughly to a suitable consistency before pelletization.This latter process aims to produce smaller pellets than are required for the finishedproduct as their volume can be increased by two to six times. The prepared pellets are fedinto the kiln which can be considered to be in three segments, namely, drying and pre-heating at the higher end and firing and then cooling at the lower end. During the progressof the prepared material through the kiln, the temperature of the clay pellets graduallyrises until expansion occurs. The expanded product is discharged from the firing zone assoon as possible for cooling to freeze the particles at the desired degree of expansion.Cooling takes place in either a rotary cooler or a fluidized bed heat-exchanger. Thefinished product is graded and, if necessary, crushed to particle sizes less than 16 mm.While the particle density can be varied depending on the range of temperatures at whichexpansion takes place, the mean expansion temperature is about 1175 ± 50°C. This variesfor different clays but most manufacturers are limited to expandable clays which have aconfined range of bloating temperatures. In some cases, the range is less than 25°Cbetween non- and full expansion. In instances such as this, the scope is limited for themanufacture of intermediate density grades. However, manufacturers of such aggregates,by requiring to optimize their production, usually have preferences for aggregates withlower particle densities, in the range 0.4–0.8 Mg/m3, as this tends to conserve the resourcewhich becomes depleted at a greater rate at higher particle densities. Thus, there is agreater attraction to produce aggregates for thermal insulation than for structural applications.However, for the lower density structural concrete, i.e. 1300–1600 kg/m3, these aggregatesare the most suitable. As shown in Figure 7.5, the surface texture is closed and smoothwith a ‘honeycombed’ or foamed internal structure, where the pores are not interconnected.

Figure 7.5 Typical expanded clay particles.

7.5.5 Expanded shale and slate aggregate

ShaleShale, a low-moisture content soft rock, is quarried, transferred to blending stockpilesbefore it is reduced by primary crushers and dry-milled to a powder of less than 250 μm.This powder is homogenized and stored ready for pelletization in manner similar to thatused for making aggregate from PFA except that no fuel is added. However, after thepellets have been produced to the appropriate size, which depends on the expansionrequired, they are compacted and coated with finely powdered limestone. The resultingpellets are spherical with a ‘green’ strength sufficient for conveying to a three-stage rotary

7/10 Lightweight aggregate manufacture

Figure 7.6 Typical expanded shale aggregate.

kiln consisting of a pre-heater, expander and cooler. Unlike other aggregates producedfrom argillaceous materials, the feedstock is reduced to a powder and then reconstitutedto form a pellet of predetermined size. The expansion (bloating) is controlled duringkilning to produce an aggregate of the required particle density. Different particle densitiesare produced by controlling the firing temperature and the rotational speed of the kiln.The coating of limestone applied to the ‘green’ pellet increases the degree of surfacevitrification which results in a particle of low permeability. This product gives versatilityto the designer for pre-selecting an appropriate concrete density. As Figure 7.6 shows,while the particle shape and surface texture of the aggregate remain essentially the same,the internal porosity can be varied according to the bloating required for the specifieddensity.

SlateSlate, a hard argillaceous rock, is first reduced to about 12–15 mm and is then fed into athree-stage kiln consisting of a pre-heater, expander and cooler which can be a pond ofwater. The firing temperature is about 1150°C at which the laminated platelets of slatebecome pyroplastic and the gases released cause the particles to expand to form an almostcubical shape. The process, which must, again, be very carefully controlled, reduces theparticle density from that of the slate, about 2.70 Mg/m3, to a pre-determined densitybetween 0.65 and 1.25 Mg/m3. The finished product is crushed and graded to a maximumsize of 25 mm and the surface texture of the finished particle is coarse and rough. However,the surface is sufficiently closed, owing to its vitrified nature, such that internally it hasan extremely low water absorption.

7.6 The future

Research on structural concrete has established that concrete made with appropriatelightweight aggregates does not have many of the deficiencies in performance comparedto concrete made with natural aggregates. As the need increases for alternatives to aggregatesfrom natural sources, there are opportunities for capitalizing on the resources that exist,for which there is currently no accepted use in their present form. For instance, it isestimated that there are some 200 million tonnes of ‘stockpiled’ PFA in landfill in the UKalone. In other EC countries landfilling of PFA is now penalized for environmentalprotection by taxation of the power companies. Although PFA has been used in the UKfor the manufacture of aggregate, the finished product has a single nominal density whichhas limited its wider application in structural concrete. This, together with the limitations

Lightweight aggregate manufacture 7/11

on the amount of excessive and variable quantities of residual fuel that the PFA sometimescontains, restricts full exploitation of the technology of conversion to a structural aggregate.

Expanded clays have their own limitations, mainly because of restrictions placed onthe availability of suitable material. In 1992 a plant for the production of expanded clayaggregate in the UK was closed due to the refusal of planning permission for the winningof otherwise accessible resources.

Other alternatives exist for combining the technologies used in the manufacture ofaggregates from both clay and PFA to produce products identical in performance to thoseof slate or shale. The amount of clay used can be as low as 10 per cent but can be as highas 80 per cent by mass when particle densities lower than 0.65 Mg/m3 are required. Theloss on ignition in the PFA is less restrictive as up to 20 per cent can be accommodatedbut it must be controlled to low variability. The advantages to manufacture are that theaggregate can be fired in a rotary kiln without a pre-heater and the cooler can be afluidized bed heat exchanger, both of which reduce the size of the plant to the essentialrotary kiln. Particle densities of the finished product can be more easily varied as therange of temperatures is controlled to produce the density required. The particle is sphericalwith a low water permeability and a smooth surface texture.

The sinter strand, can use any argillaceous material, for instance clay or the considerablestockpiles of ‘soft’ slate that was produced as waste when quarried (e.g. in North Wales,while the spoil and tailings produced from mining coal are, in theory, potential rawmaterials for lightweight aggregate production). However, the volatiles in mine tailings,etc. present a concern for air pollution control while clay, shales and slate containing lowamounts of volatile products are most suitable. However, for lightweight aggregates to besuccessful in the future, pulverization and reconstitution of the shale or slate into pelletsmust be an integral part of the manufacturing process. Producing aggregates on a sinterstrand results in a less spherical particle but with the advantage of a closed surface.

For increasing the suitability of mine ‘tailings’ as an aggregate feedstock, the removalof volatiles such as naphtha, sulfur and bituminous by-products would be essential beforepulverization and homogenization of the resultant ‘clean’ feedstock prior to pelletization.

The most significant development is the use of resources from so-called ‘waste’ materialsfor the manufacture of lightweight aggregates (Cheeseman et al., 2000). Such resourcematerials include residues from the incineration of municipal waste (incinerator bottomash (IBA) and air pollution control residues), sewage sludge, waste clay, etc. This willaddress two major problems, namely, the increasing difficulty of waste disposal due toEU restrictions on landfill and the provision of aggregate which is becoming more difficultto obtain for a variety of reasons, particularly in conurbations where there is reducedaccess to local resources. The use of state-of-the-art efficient thermal processing techniquesand the financial benefits resulting from the use of ‘wastes’ will allow the price of themanufactured aggregate to be reduced to a level at which lightweight concrete of a givengrade will be no more, and possibly less, expensive than normal dense concrete but withenhanced properties (Owens and Newman, 1999).

7.7 Conclusions

The technology is available for successful lightweight aggregate manufacture. However,to produce a generic and consistent aggregate the feed stocks used must be statistically

7/12 Lightweight aggregate manufacture

homogenized before pelletization. The manufacture of aggregates in the UK from slatewaste in North Wales and from minestone in East Kent, Derbyshire and Scotland, allfailed, not because of technical deficiencies of the aggregate itself, but because of difficultieswith the preparation of an inconsistent raw material, its process control and emissionsfrom the plants. Alternatively, aggregates have been produced from slag, a highly controlledby-product of iron production, as well as from PFA. Both have succeeded since considerablecare is taken to remove variability from the raw materials to enable production to continuewith as little ‘manual’ interference as possible.

Another concern with all lightweight aggregate manufacture is to match demand withsupply. Significant investment will not be made until it becomes common practice to uselightweight concrete for structural purposes both below, as well as above, ground level.For example, One Shell Plaza in Houston (Khan, 1969) used lightweight concrete in thefoundations and basement, proving that engineers can not only utilize the advantages oflightweight concrete, but can also make it commercially advantageous for owners of alltypes of structures.

References

BS EN 206-1: Concrete: Part 1: Specification, performance, production and conformity.BS EN 13055: Part 1: Lightweight aggregates: Part 1: Lightweight aggregates for concrete, mortar

and grout.BS EN 1097-3, Tests for mechanical and physical properties of aggregates: Part 3: Determination

of loose bulk density and voids.BS EN 1097-6, Tests for mechanical and physical properties of aggregates: Part 6: Determination

of particle density and water absorption (Annex C).BS EN 13055: Part 1: Lightweight aggregates: Part 1: Lightweight aggregates for concrete, mortar

and grout (Annex A: Determination of crushing resistance).Cheeseman, C., Newman, J. and Owens, P. (2000) Why waste waste? A blueprint for generating

products from waste in London. Institute of Wastes Management, Millennium Competition,Award Winning Entries, IWM, October.

Concrete Society (1978) Structural Lightweight Aggregate Concrete for Marine and Off-shoreApplications, Report of a Concrete Society Working Party, Technical Report No. 16, May.

Cripwell, J.B. (1992) What is PFA? Concrete, 26, No. 3; May/June, 11–13.Khan, F.R. (1969) Lightweight concrete for total design of One Shell Plaza. ACI Special Publication

SP29, Lightweight Concrete Paper SP29-1.Owens, P.L. and Newman, J.B. (1999) Increasing the environmental acceptability of new energy

from waste plants by integration with cost effective concrete aggregate manufacture. IWA Scientific& Technical Review, Institute of Wastes Management, Nov. 21–26.

RILEM (1976) Functional classification of lightweight concretes, Recommendation LC2, 2nd edn.Rudnai, G. (1963) Lightweight Concretes. Akademiar Kiado Publishing House of the Hungarian

Academy of Sciences, Budapest.

8

The effects of naturalaggregates on the

properties of concreteJohn Lay

8.1 Aims and objectives

The purpose of this chapter is to explain how the properties of natural aggregates influencethe properties of fresh and hardened concrete. It describes British and European Standardtests for aggregates and emphasizes the importance of correct sampling and describes theinterpretation of aggregate test results. The chapter also considers the likely effects ofimpurities associated with natural aggregates.

8.2 Brief history

One of the earliest known concretes was placed in 5600 BC on the banks of the riverDanube (Stanley, 1979). It contained sand and gravel aggregate, bound with lime. Ancientconcretes have also been found consisting of broken rock, sand and lime. For manycenturies, transport constrained the choice of aggregates. In almost all cases, local materialswere used. Even so, differences between the performance of different rocks in concretehave been known since Roman times (Sims and Brown, 1996). In the modern age, it hasbecome more practical to transport aggregates and specifications have grown up to assess

8/2 The effects of natural aggregates on the properties of concrete

materials. Early specifications contained subjective clauses, such as the requirement thataggregates should be clean, hard and durable. The phrase survives to this day in theSpecification for Highway Works (Highways Agency, 1998). In the latest version of BS882, the British Standard Specification for Aggregates from Natural Sources for Concrete,the requirements are expressed quantitatively and concisely (BSI, 1992). BS EN 12620,the European Standard for Aggregates for Concrete, also defines requirements quantitatively(CEN, 2002).

8.3 Introduction

Aggregate is the main constituent of concrete. Aggregate properties do influence concreteproperties, but by and large they do not control the performance of the concrete. Theessential requirement is that the aggregate remains stable within the concrete in its exposureconditions. The choice of aggregates used is often constrained by haulage costs.

There are three main reasons for mixing aggregate with cement paste to form concrete,rather than using cement paste alone. The first and oldest reason is that aggregate ischeaper than cement, so its use extends the mix and reduces costs. Second, aggregatereduces shrinkage and creep, giving better volume stability. Third, aggregate gives greaterdurability to concrete. Many deterioration processes principally affect the cement paste.

There are clear economic and technical reasons for using as much aggregate and aslittle cement as possible in a concrete mix. Since the economic arguments are obvious,the rest of this chapter will consider technical aspects of the relationship between theproperties of aggregate and the properties of concrete.

8.4 Classification

Petrological classification alone is an inadequate guide to aggregate performance. Whilepetrographic descriptions of aggregates are interesting and useful in assessing quality,they do not provide any more than general guidance on performance in concrete. TheDoE mix design method, Design of normal concrete mixes (Teychenné et al., 1988), takesinto account only two types of aggregate: crushed and uncrushed. The nominal maximumsize of the coarse aggregate is considered. Fine aggregate is characterized by the proportionpassing a 600 μm test sieve. The particle density of the combined aggregate is alsorequired.

Mix design methods based on trial mixing are likely also to take account of geologicaltype, bulk density (see section 1.10) and 10 per cent fines value (see section 1.9).

8.5 Sampling

Although it is not an aggregate property, correct sampling, sample reduction and samplepreparation are essential to the determination of all aggregate properties. The results oftests are meaningless if sampling procedures are not strictly observed. The aim of samplingis to obtain a test portion test that is representative of the average quality of the batch

The effects of natural aggregates on the properties of concrete 8/3

(BSI, 1989a). A batch is a definite quantity of aggregate produced under conditions whichare presumed to be uniform (BSI, 1989b).

British and European Standards give detailed procedures for obtaining and reducingsamples (BSI, 1989a; CEN, 1996, 1999). The process for obtaining samples involvestaking increments at random from the batch of aggregate and combining the incrementsto form a bulk sample. Table 8.1 shows the minimum number of increments required toform a bulk sample according to the requirements of BS 812: Part 102. The EuropeanStandard, BS EN 932-1, requires that the quantity of sampling increments taken to formthe bulk sample be based on previous experience. Research funded by the EuropeanUnion indicates that a minimum of fifteen increments is necessary in most cases (Ballmannet al, 1996). BS EN 932-1 recommends that the minimum mass of bulk sample should besix times the square root of the maximum particle size (in mm) multiplied by the loosebulk density (in Mg/m3). This leads to similar bulk sample masses to those shown inTable 8.1.

Table 8.1 Minimum number of sampling increments (BSI, 1989a)

Nominal size of Minimum number of sampling increments Approximate minimumaggregate —————————————————— mass for normal density(mm) Large scoop Small scoop aggregate

(at least 2 litres) (at least 1 litre) (kg)

28 and larger 20 – 505 to 28 10 – 255 and smaller 10 half scoops 10 10

Sample reduction is usually necessary to provide a laboratory sample from the bulksample, and then test portions from the laboratory sample. The principle is again toensure that the reduced sample is representative of the larger sample, and thereforerepresentative of the batch. Quartering and riffling are suitable methods for sample reduction.

Quartering (which in fact halves the sample) involves mixing the sample and dividingit into approximately equal quarters. A pair of diagonally opposite quarters is retainedand combined to form the reduced sample. The process can then be repeated to furtherreduce the sample mass if necessary.

Riffling involves passing the sample through a sample divider consisting of a row ofchannels. Alternate channels lead to two separate boxes. When the sample is passedthrough the channels, it is separated into two halves, one half being retained in each box(Figure 8.1).

For some tests, a further sample preparation stage is required. For example, a numberof tests for mechanical properties are carried out on the fraction of the aggregate passinga 14 mm test sieve but retained on a 10 mm test sieve. For some chemical tests, thesample is ground to a fine powder and further reduced before analysis. Details of thesefinal sample preparations are included in the test procedures.

8.6 Grading

Grading is the common term used to mean the particle size distribution of the aggregate.Surprisingly, variations in grading have only minor effects on hardened concrete properties.

8/4 The effects of natural aggregates on the properties of concrete

Figure 8.2 (Miller, 1978) shows the effect of changes in a coarse aggregate grading oncompressive strength at constant workability. All concrete was batched to 50 mm slump.It is interesting to note that the BS 882 limits for the percentage passing the 10 mm sieve

Figure 8.1 Methods of sample reduction.

(a) Quartering using a palette knife (b) Quartering using a quartering cross

(c) A riffle box sample divider

80

60

40

Com

pres

sive

str

engt

h (M

Pa)

a/c ratio

4.5

6.0

0 10 20 30 40 50 60 70 80 90 100Percentage passing 10 mm test sieve in 20–5 mm aggregate

(same shape material)

Figure 8.2 The effect of coarse aggregate grading on compressive strength.

The effects of natural aggregates on the properties of concrete 8/5

in 20–5 mm graded aggregate are 30–60 per cent. The graphs are virtually flat lines inthis range. Figure 8.3 (Ryle, 1988) shows the effect of changes in sand grading at constantworkability. Within the typical range of sand gradings, the effect is fairly small. Again, itis interesting to note that the peaks of the curves fall more or less at the mid-point of theoverall BS 882 limits for the percentage of sand passing the 600 μm test sieve. Theinfluence of grading on fresh concrete is much greater. Workability, cohesion, handling,compaction, bleeding and finishing can all be affected. Grading limits have usually beenderived to maintain consistency with available materials.

Figure 8.3 The effect of sand grading on compressive strength.

35% sand

40% sand

45% sand

20 40 60 80 100Percentage passing 0.6 mm sieve

38

36

34

32

30

28

Com

pres

sive

str

engt

h (M

Pa)

The theoretical objective of aggregate grading is to provide the maximum relativevolume occupied by the aggregate, while maintaining the lowest aggregate surface area.This is because the space in the concrete mix not occupied by aggregate must be filledwith cement paste, and because the surface of all the solids in the mix must be wetted. Inpractice, a compromise between these demands, which in some ways conflict, must beachieved. Modern mix design methods mean that aggregates with quite a range of gradingsmay be used (Dewar, 1983). However, uniformity and consistency of materials are stillneeded.

Particle size distribution is determined by drying test portions of aggregate and passingthem through a nest of sieves of decreasing aperture (BSI, 1985). The amount of materialretained on each sieve is weighed and the masses used to calculate the percentage of thesample that has passed each test sieve. It is preferable to wash the test portions free offines and redry them before sieving. This breaks down any agglomerated particles in thesample (BSI, 1985). It is important not to overload the test sieves because they may blindand retain particles finer than the sieve aperture.

Grading limits in British Standards are specified in the form of tables. For eachaggregate product, test sieves are listed alongside allowable ranges for the percentage ofmaterial passing that sieve. In European Standards, the aggregate producer is required todeclare a grading and remain within a specified tolerance of the declared grading. Thereare also some broad overall limits for products. The European Standard approach isintended to encourage consistency.

8/6 The effects of natural aggregates on the properties of concrete

8.7 Maximum size of aggregate

From the point of view of economy and minimum surface area, it would appear that thelarger the maximum size of the aggregate, the better. However, larger aggregate sizesintroduce problems in handling and placement, especially in reinforced concrete. Largeparticles may be unable to pass between congested reinforcement bars or may exceeddepth of cover to reinforcement. Another problem with larger materials is that, while thereduced surface area needs less wetting, it also provides less area for the paste to bond tothe aggregate. This can lead to reduced cohesion. It also results in a higher bond stress,and may lead to weaker concrete. Larger particles of crushed rock are far more likely tocontain defects and will often be weaker than smaller particles of the same material(Neville, 1995). Large sizes are rare in gravel deposits. In general, lean mixes benefitfrom a large aggregate size (up to 150 mm if available). For normal strength concretes,optimum maximum size usually lies between 20 mm and 40 mm. For very high strengthconcretes, 10 mm aggregate is best, probably due to improved bond (Neville, 1995). (Seevolume 3, Chapter 3 in this series for more detail.)

8.8 Aggregate shape and surface texture

The shape of aggregate particles can affect water demand and workability, mobility,bleeding, finishability and strength. Equidimensional shapes are preferred. Flaky or elongatedparticles tend to be detrimental. Rounded particles tend to give better workability thancrushed or angular particles. Producers can exercise some control over particle shapethrough the processing of the aggregate.

The surface texture of aggregate particles can range from glassy, through smooth,granular, rough and crystalline, to honeycombed. The texture is only really an issuewhere flexural strength is important, or for very high-strength concretes. In both cases,rougher textures give greater strengths, all other things being equal, because the aggregate–cement paste bond is improved.

Figure 8.4 (Ryle, 1988) shows the relationship between cement content and compressivestrength for a crushed rock, a land-won flint and a marine flint. At higher cement contents,the angular and rough crushed rock produces a higher-strength concrete because of improvedbond. At low- and medium-cement contents, the smooth and rounded marine gravelexerts a lower water demand for the same workability than crushed rock because there isless surface to wet.

The most common shape test is flakiness index. Coarse aggregate particles are presentedto a special gauge (Figure 8.5). If a particle’s least dimension is less than 60 per cent ofits mean dimension, it passes through the gauge and is classed as flaky (BSI, 1989b).Flakiness index is defined as the percentage by mass of flaky particles in the sample. TheEuropean Standard flakiness test operates on a similar principle, using sieves with rectangular(rather than square) apertures. Confusingly, this test classes a particle as flaky if its leastdimension is less than 50 per cent of its upper sieve size. A test survey has established anapproximate relationship between the two indices (Eurochip, 1996). The survey showsthat the British Standard requirements of maximum flakiness index 50 for uncrushedgravel and 40 for crushed rock and crushed gravel are equivalent to European Standardcategories of 50 and 35 respectively.

The effects of natural aggregates on the properties of concrete 8/7

Figure 8.4 A relationship between cement content and compressive strength.

100 200 300 400 500

Cement content (kg/m3)

Crushed rock

Land flint

Marine flint

80

70

60

50

40

30

20

10

0

Com

pres

sive

str

engt

h (M

Pa)

Figure 8.5 Flakiness and elongation.

(a)

(b)

Flaky particles (left)

Elongated particles(right)

Elongation gauge(above)

Flakiness gauge(below)

8/8 The effects of natural aggregates on the properties of concrete

Other shape tests include elongation index and shape index. A particle is classed aselongated (using a gauge) if its longest dimension is greater than 180 per cent of its meandimension. No European Standard for elongation index is planned. Shape index is definedas the proportion of particles in a sample whose greatest dimension is more than threetimes their least dimension. A slide gauge is used to assess particles. A European Standardtest for shape index exists, but flakiness index will be the reference method for thedetermination of shape.

Where an assessment of surface texture is required, a qualitative description is usuallygiven. Quantitative methods tend to be laborious and not widely used.

8.9 Aggregate strength

Aggregate strength is rarely a problem. The strength of concrete cannot exceed thestrength of the aggregate. In practice, concrete grade is likely to be much less than thestrength of the aggregate, because stress concentrations are generated when the concretestresses are shared by the aggregate and cement paste. Weak aggregates may break downduring mixing, handling and compaction.

Within British Standards, the mechanical properties of aggregate can be assessed from10 per cent fines value or aggregate impact value. In the 10 per cent fines test, a load isapplied to the 14–10 mm fraction of the aggregate. The load required to partially crushthe sample so that 10 per cent of it passes a 2.36 mm test sieve is called the 10 per centfines value. In the aggregate impact value test, the 14–10 mm fraction of the aggregatereceives fifteen blows from a drop hammer. The aggregate impact value is the proportionof the sample that passes a 2.36 mm sieve after the hammer impacts.

European Standards will not contain the 10 per cent fines test or the aggregate impactvalue test (although a different impact test, the German Schlagversuch test, has beenincluded). The reference mechanical test will be the Los Angeles test. The 14–10 mmfraction of the aggregate is placed in a steel drum with eleven large steel balls. The drumis rotated 500 times, partially fragmenting the sample. The Los Angeles value is theproportion of the sample passing a 1.6 mm sieve after fragmentation.

There is no direct correlation between 10 per cent fines value or aggregate impactvalue and Los Angeles value. BS 882, The British Standard Specification for Aggregatesfrom Natural Sources for Concrete, specified 10 per cent fines values or aggregate impactvalues based upon the end use of the concrete. Because the Los Angeles value does notcorrelate with the required properties, the UK National Guidance (BSI, 2003) to 12620,Aggregates for Concrete, recommends a single value for normal concrete (Table 8.2).

8.10 Aggregate density

Loose bulk density reflects the grading and shape of aggregate, since it measures theproportions of voids in a given volume of aggregate. Its main use is as a consistencycheck on materials. It is also used to convert between mix proportions by mass and byvolume (BSI, 1995).

The effects of natural aggregates on the properties of concrete 8/9

Particle density is largely a consequence of geological type. Aggregate producers canhave some influence, for example by processing to remove lightweight particles with aparticle density of more than 2000 kg m–3 (ASTM, 1994). Particle density affects concreteyield and density.

Water absorption is usually determined at the same time as particle density. Absorptionaffects mix design and density, and is indirectly related to shrinkage, soundness, frostsusceptibility and general durability (Smith and Collis, 1993). Absorption limits are rare,though may be useful if a link to an undesirable concrete property such as frost susceptibilityis established, because absorption is easily determined.

Bulk density is determined by filling a container of known volume with aggregate. Themass of aggregate required to fill the container divided by the volume gives the bulkdensity. The value is usually determined as loose bulk density, that is, the aggregate isuncompacted. The British Standard describes and the European Standard allows thedetermination of bulk density in other, compacted conditions.

Particle density and absorption are determined by saturating the aggregate and weighingit in a saturated and surface dry condition and an oven-dried condition. Volume is determinedfrom water displacement.

8.11 Drying shrinkage

The drying shrinkage of concrete is usually due to movement of the cement paste, becausemost natural aggregates do not undergo any appreciable drying shrinkage (BRE, 1991).Indeed, most aggregates restrain concrete shrinkage because they are less elastic than thecement paste to which they are bonded. Concretes with higher aggregate contents shrinksubstantially less than cement-rich mixes all else being equal.

Some rock types undergo unusually large volume changes on wetting and drying(BRE, 1991). These rocks are typically weathered, and contain clay or mica minerals.The absorption of the rock is sometimes high (Smith and Collis, 1993). If processed intoaggregates, these rocks produce concrete with poor volume stability and a tendency todeflect and crack.

The British and European Standard tests for drying shrinkage of aggregates are similar.Concrete prisms are manufactured using the aggregates under test. The mix proportionsof the concrete are fixed. The prisms are cured in water for five days and their length ismeasured. The prisms are then oven-dried and measured again. The difference betweenthe two measurements expressed as a percentage of the final prism length is the dryingshrinkage.

Table 8.2 Limits on the mechanical properties of aggregates

Type of concrete BS 882 requirements EN 12620 category recommended————————————————— for United KingdomTen per cent fines Aggregate impact Los Angeles valuevalue (minimum) value (maximum) (maximum)kN % %

Heavy-duty concretefloor finishes 150 25 40Pavement wearing surfaces 100 30 40Others 50 45 40

8/10 The effects of natural aggregates on the properties of concrete

Aggregates giving a concrete drying shrinkage of 0.075 per cent or less are suitable forall concreting purposes. Aggregates with higher shrinkages are only suitable for concretewhere complete drying out never occurs, for mass concrete surfaced with air-entrainedconcrete or for symmetrically and heavily reinforced members not exposed to the weather(BRE, 1991).

8.12 Soundness

A sound aggregate is able to resist stresses induced by environmental or climatic conditions.Some aggregates degrade or disintegrate under the action of cycles of freezing and thawingor salt weathering. Susceptible aggregates are usually microporous, that is, they havehigh absorption arising from a large proportion of fine pores (Neville, 1995; Sims andBrown, 1996). Freezing or salt crystallization can induce considerable stresses in finepores that may lead to degradation. Unfortunately, there is no definite correlation betweenthe soundness of aggregate and its performance in concrete. This is partly due, for example,to the resistance of concrete to cycles of freezing and thawing being affected by factorssuch as the degree of saturation of the concrete, its maturity, its quality and the severityof its exposure conditions as well as aggregate soundness.

British and European Standards contain the magnesium sulfate soundness test (Figure8.6). In this test, the 14–10 mm fraction of the aggregate is placed in a wire basket andsoaked in saturated magnesium sulfate solution. The sample is oven-dried to crystallizemagnesium sulfate in the aggregate pores. The soaking and drying is repeated five times.The magnesium sulfate soundness value is defined in British Standards as the percentageof the sample retained on the 10 mm sieve after test. In European (and ASTM) Standards

Figure 8.6 Aggregate soaking in saturated magnesium sulfate solution during soundness testing.

The effects of natural aggregates on the properties of concrete 8/11

the value is the percentage passing the 10 mm sieve after test. The procedure can bemodified to examine other aggregate fractions, replacing the 10 mm sieve with the appropriatelower size sieve for the fraction.

A European Standard test subjecting aggregates to cycles of freezing and thawing isalso available. There is no experience of this test in the United Kingdom.

8.13 Thermal properties

The thermal coefficient of expansion of aggregates typically ranges from 2 to 16 micro-strain/°C in normal temperature ranges (Harrison, 1992). The coefficient for silica mineralsis higher than that for most other minerals. The expansion coefficient of an aggregate istherefore broadly related to its silica content, but wide variations can occur. Limestone isusually low in silica and has a low coefficient, so it may be specified for concrete whena particular risk of early-age thermal cracking has been identified.

The expansion coefficient of cement paste is higher than that of aggregate. Becauseconcrete consists mainly of aggregate, the thermal coefficient of expansion of concrete islargely determined by the coefficient of the aggregate and the aggregate content. Limestoneis the preferred aggregate for concrete when the expansion coefficient of the concreteneeds to be kept low. While aggregate type is important, mix design, moisture conditionand the age of the concrete are also significant influences. For these reasons, it is difficultto produce concrete with a thermal coefficient of expansion of less than 8–10 micro-strain/°C.

All aggregates produce fire-resistant concrete and the degree of fire resistance isrelated to aggregate type. Aggregates containing high levels of quartz offer poorer resistance,because quartz undergoes an expansive solid phase change at high temperature (Neville,1995). Limestone contains little or no quartz and therefore offers better fire resistance.Lightweight aggregates are usually stable at high temperatures, have low coefficients ofthermal expansion and provide heat insulation. They produce concrete with very goodfire resistance (for more information see Chapter 5, Volume 2 in this series).

Thermal coefficient of expansion can be determined using strain gauges and atemperature-controlled cabinet. The change in the strain of the aggregate is measured astemperature changes and is used to calculate thermal coefficient of expansion (US Army,1963).

8.14 Fines content

Fines are defined in British Standards as materials passing a 75 μm test sieve. Europeanstandards draw a distinction between sand and fines at 63 μm. This material may be veryfine sand, silt, dust or clay. A moderate fines content fulfils a void-filling role, and aidscohesion and finishability. Excessive fines, especially fines consisting of clay, increasewater demand and reduce the aggregate–cement paste bond. Both of these effects resultin reduced strength.

Fines content is usually determined in conjunction with particle size distribution. Thedried sample is washed over a guarded 75 μm (or 63 μm) test sieve (guarded with a nested

8/12 The effects of natural aggregates on the properties of concrete

1.18 mm test sieve to prevent coarser particles damaging the 75 μm sieve) and thenredried. The loss in mass is equivalent to the fines content.

Generally, more fines can be tolerated in crushed rock than in sand and gravel, becausethe fines consist of rock powder rather than clay or silt. BS 882 limits and recommendedEN 12620 categories are shown in Table 8.3.

Table 8.3 Limits for fines content

Aggregate type BS 882 limit EN 12620 category recommendedfor United Kingdom

(max % passing 75 μm) (max % passing 63 μm)

Coarse gravel 2 1.5Coarse crushed rock 4 4Gravel sand 4 3Crushed rock sand 16 16Crushed rock sand inheavy duty floor finishes 9 10All-in gravel 3 3All-in crushed rock 11 11

8.15 Impurities

So far, this chapter has considered properties of the aggregate relating to the intrinsicproperties of the mineral deposit and the characteristics of the processed product. Mostaggregates also contain traces of materials that are not aggregate. These materials arecalled impurities and generally impart no benefit to concrete manufactured using theaggregate. Chlorides, sulfates and shell will be considered briefly. The effects of someother impurities are tabulated.

Chlorides occur in marine aggregates and some coastal and estuarine deposits. Processingremoves the majority of chloride, but a residual amount usually remains in marine aggregateproducts. High levels of chloride can accelerate the set of fresh concrete and can lead todamp patches and efflorescence on hardened concrete as the salt migrates to the surface.However, the main problem with chlorides is that their presence increases the risk ofcorrosion of embedded metal. The risk of corrosion is related to the chloride ion contentof the concrete expressed as a proportion of the cement content. It is not, therefore,possible to specify generally how much chloride can be tolerated in aggregate. Butbecause aggregate constitutes such a large proportion of concrete, even relatively smallpercentages of chloride in the aggregate can contribute fairly large amounts of chlorideion to the concrete mix. For this reason, guideline limits on chloride content of aggregatestend to be low (BS, 1992). Guidelines are related to cement type and end use of theconcrete because the risk and possible consequences of corrosion depend upon thesefactors. BS 882 guidelines are shown in Table 8.4.

Sulfates are found in natural ground, usually associated with ancient clays (ThaumasiteExpert Group, 1999). On rare occasions these sulfates may be present as impurities inaggregates. High levels of sulfate can interfere with cement hydration in fresh concretebut the major concern is ‘normal’ sulfate attack of hardened concrete. (Other forms ofsulfate attack are described in Volume 2, Chapter 12 in this series. In wet conditions,sulfates can react with components of Portland cement. The process is expansive and may

The effects of natural aggregates on the properties of concrete 8/13

lead to cracking and spalling of concrete. The risk of sulfate attack is related to the sulfatecontent of the concrete expressed as a proportion of the cement content. BS 882 requiresproducers to provide information on sulfate content of aggregates when required, toallow concrete producers to assess the amount of sulfate contributed to the mix by theaggregates. BS EN 12620 limits the sulfate content of natural aggregates to 0.2 per cent.

Shell is found to a greater or lesser extent in most marine aggregates. High levels ofshell in aggregate can reduce the workability of fresh concrete and can make concretemore difficult to finish. The surface finish of hardened concrete may be affected by shell,and there is some evidence that hollow shells near the concrete surface may make concretemore susceptible to freeze–thaw damage (Lees, 1987). BS 882 limits the shell content of10 mm aggregate to 20 per cent and coarser aggregates to 10 per cent. BS EN 12620limits the shell content of all coarse aggregates to 10 per cent. Neither Standard limits theamount of shell in fine aggregate (see also Table 8.5).

Table 8.4 Guidance on the chloride content of aggregates

Type and use of concrete Maximum chloride ion content expressed as percentageby mass of combined aggregate

Prestressed concrete and heat-cured 0.01concrete containing embedded metal

Concrete containing embedded metal 0.03made with sulfate-resisting cement

Concrete containing embedded metal 0.05made with cements or combinationsother than sulfate-resisting cement

Other concrete No limit

Table 8.5 Some other impurities and their effects

Impurity Effect on fresh concrete Effect on hardened concrete

Acid soluble material in sand None Reduced skid resistance of pavementquality concrete

Alkali reactive silica None Risk of alkali–silica reactionSwelling clays Increased water demand Reduced strengthReactive iron pyrites Possible reduced yield Surface stainingMica Increased water demand Reduced strengthOrganic matter Possible retardation Possible reduced strengthCoal and lignite Possible retardation Surface staining, pop-outsSoluble lead or zinc Possible retardation Possible reduced strength

8.16 Summary

Aggregate is the main constituent of concrete and the properties of aggregate influencethe properties of concrete. Petrological classification alone is an inadequate guide toaggregate performance. Any determination of aggregate properties requires samples to betaken. Correct sampling, sample reduction and sample preparation are essential if testresults are to be meaningful.

8/14 The effects of natural aggregates on the properties of concrete

Consistency of particle size distribution is important to the properties of fresh concrete.There is a more minor influence on hardened concrete properties. Larger maximumparticle size benefits lean concrete but the optimum size is around 20 mm for normalstrength concrete. Equidimensional particle shapes are best. Shape and texture affectwater demand and aggregate–cement paste bond. Aggregate must be strong enough towithstand mixing, handling and compaction without degradation. In practice, the aggregatestrength usually exceeds concrete strength considerably. Bulk density can be used as acheck on the consistency of aggregate grading and shape. Particle density and absorptionmust be known for mix design purposes.

Most aggregates restrain concrete shrinkage. A few aggregates undergo large volumechanges on wetting and drying, and may produce concrete with poor volume stability.Some aggregates are less able to cope than others with stresses induced by environmentalconditions. Susceptible aggregates may disintegrate after cycles of freezing and thawing.

The thermal coefficient of expansion of concrete is largely determined by the aggregate.All aggregates produce fire-resistant concrete though lightweight aggregates are usuallybest.

Moderate fines contents are useful for cohesion and finishability of concrete. Excessivefines contents increase water demand and may interfere with the aggregate–cement pastebond. Occasionally, impurities are associated with aggregates. At best, impurities impartno benefit to concrete. If present in sufficient quantities, their effects can be detrimental.Aggregate is essential to the economy, stability and durability of concrete.

References

American Society for Testing and Materials (1994) C123 – Standard Test Method for LightweightPieces in Aggregate. ASTM, West Conshohocken.

Ballmann, P., Collins, R., Delalande, G., Mishellany, A. and Sym, R. (1996) Determination ofthe number of sampling increments necessary to make a bulk sample for tests for aggregates.European Community Measurement and Testing Programme.

BRE Digest 357 (1991) Shrinkage of natural aggregates in concrete. Building ResearchEstablishment: Watford.

British Standards Institution (1985) BS 812 – Testing aggregates. Part 103 – Methods fordetermination of particle size distribution – Section 103.1 – Sieve tests. BSI, London.

British Standards Institution (1989a) BS 812 – Testing Aggregates. Part 102 – Methods forsampling BSI, London.

British Standards Institution (1989b) BS 812 – Testing Aggregates. Part 105 – Methods fordetermination of particle shape – Section 105.1 – Flakiness index. BSI, London.

British Standards Institution (1992) BS 882 – Specification for Aggregates from natural sourcesfor concrete. BSI, London.

British Standards Institution (1995) BS 812 – Testing Aggregates. Part 2 – Methods of determinationof density. BSI, London.

British Standards Institution (2003) PD 6682 – Aggregates – Part 1: Aggregates for concrete –Guidance on the use of BS EN 12620. BSI, London.

Dewar, J D. (1983) Computerized simulation of aggregate, mortar and concrete mixtures. ERMCOCongress: London.

Eurochip (1996) Project report. Buckinghamshire Highways Laboratory Planning andTransportation Department.

European Committee for Standardization (CEN) (1996) BS EN 932 – Tests for general propertiesof aggregates. Part 1. Methods of sampling. BSI, London.

The effects of natural aggregates on the properties of concrete 8/15

European Committee for Standardization (CEN) (1999) BS EN 932 – Tests for general propertiesof aggregates. Part 2. Methods of reducing laboratory samples. BSI, London.

European Committee for Standardization (CEN) (2002) BS EN 12620 Aggregate for concrete, BSI,London.

Harrison, T.A. (1992) CIRIA Report 91: Early-age thermal crack control in concrete (revisededition). Construction Industry Research and Information Association, London.

Highways Agency (1998) Manual of Contract Documents for Highway Works. Volume 1:Specification for Highway Works. The Stationery Office: London.

Lees, T.P. (1987) C&CA Guide: Impurities in concreting aggregates. Cement and ConcreteAssociation, Slough.

Miller, E.W. (1978) The effect of coarse aggregate grading on the properties of fresh and hardenedconcrete. Institute of Concrete Technology. Advanced Concrete Technology Project,Crowthorne.

Neville, A.M. (1995) Properties of Concrete (4th edn). Longman, Harlow.Ryle, R. (1988) Technical aspects of aggregates for concrete. Quarry Management, April.Sims, I. and Brown, B.V. (1996) Concrete aggregates. In Hewlett, P.C. (ed.), Lea’s Chemistry of

Cement and Concrete (4th edn). Arnold, London, 903–1011.Smith, M.R. and Collis, L. (1993) Geological Society Engineering Geology Special Publication

No 9. Aggregates. Sand, gravel and crushed rock for construction purposes (2nd edn). TheGeological Society, London.

Stanley, C.C. (1979) Highlights in the History of Concrete. Cement and Concrete Association,Slough.

Teychenné, D.C., Franklin, R.E., Erntroy, H.C., Nicholls, J.C. and Hobbs, D.W. (1988) Designof normal concrete mixes. Department of the Environment, London.

Thaumasite Expert Group (1999) The thaumasite form of sulfate attack: Risks, diagnosis, remedialworks and new construction. Department of the Environment, Transport and the Regions, London.

United States Army (1963) CRD–C 125–63, Method of test for coefficient of linear thermal expansionof coarse aggregate (strain-gage method). US Army, Vicksburg, Miss.

Further reading

Geological Society Engineering Geology Special Publication No. 9, Aggregates, is a comprehensiveguide to aggregates for all construction purposes.

Lea’s Chemistry of Cement and Concrete (4th edn) contains a substantial chapter on concreteaggregates.

Appendix Relevant standards for natural aggregates

British Standards

BS 812 Testing aggregates. A standard consisting of twenty parts, givingmethods for sampling, examining and testing aggregates

BS 882 Specification for aggregates from natural sources for concreteBS 1199/1200 Specification for building sands from natural sources

European Standards

EN 932 Tests for general properties of aggregates. In six parts, referring tosampling, petrography, equipment and precision.

8/16 The effects of natural aggregates on the properties of concrete

BS EN 933 Tests for geometrical properties of aggregates. In ten parts, coveringparticle size distribution, shape, texture, shell and fines.

BS EN 1097 Tests for mechanical and physical properties of aggregates. In ten parts,covering strength, density, absorption, polishing and abrasion.

BS EN 1367 Tests for thermal and weathering properties of aggregates. In five parts,covering freeze–thaw, volume stability and heat.

BS EN 1744 Tests for chemical properties of aggregates. In four parts, coveringchemical analysis and alkali–silica reaction.

BS EN 12620 Aggregates for concrete.

Limits for aggregate properties in BS EN 12620 are generally expressed as a range ofvalues related to categories. This allows specifiers to choose limits suitable for the materials,climate, uses and practice in their region of Europe. A Published Document (BSI, 2003)provides guidance and recommends categories for use in the UK.

Index

Abrasion and erosion, silica fume basedconcrete, 3/45

Abrasion resistance, CACs, 2/24–5Accelerating admixtures:

about accelerating admixtures, 4/20–1applications, 4/21calcium chloride in, 4/20chloride-free, 4/20mechanism, 4/21performance, 4/21see also Admixtures

Acid/sewage resistance:CAC concrete, 2/23–4fly ash in concrete, 3/17metakaolin based concrete, 3/55

Admixtures:about admixtures, 4/3–4anti-bleed admixtures, 4/34benefits, 4/5–6CAA (Cement Admixtures Association),

4/34cement/concrete chemistry aspects, 4/7–8definition, 4/3dispensers, 4/35EFCA (European Federation of Concrete

Admixture Associations), 4/34foamed concrete/CLSM, 4/31–2grout admixtures, 4/34health and safety, 4/36history, 4/4mechanism of action, 4/5–6

mortar admixtures, 4/33polymer dispersions, 4/32pre-cast, semi-dry admixtures, 4/32primary and secondary properties, 4/6pumping aids, 4/30rheology/rheology measurements, 4/6–7sprayed concrete admixtures, 4/31standards, 4/4–5storage, 4/35supply/suppliers, 4/34time of addition, 4/35truck washwater admixtures, 4/32–3types without European standards, 4/5see also Accelerating admixtures; Air-

entraining admixtures; Corrosion-inhibiting admixtures; Dispersingadmixtures; Retarding admixtures;Shrinkage-reducing admixtures;Underwater/anti-washout admixtures;Water resisting (waterproofing)admixtures; Water-reducing admixtures

Aggregate classification and sources:about Aggregates, 5/3–4, 8/13–14classification, 5/21–3, 8/2grading, 8/3–5quarry assessment, 5/23–5quartering, 8/3, 8/4riffling, 8/3, 8/4sample reduction, 8/3, 8/4sampling, 8/2–3sources and types, 5/12–21

I/2 Index

UK locations, 5/15see also Crushed rock; Lightweight aggregate;

Limestone; Rocks; Sands and gravelsAggregate impurities:

about deleterious materials, 5/25–6chalk particles, 5/33chlorides, 8/12clay and weathered rocks, 5/30–1coal particles, 5/33–4concrete setting problems, 5/26–8effects on concrete, 8/12–13lignite particles, 5/33–4micas, biotite and muscovite, 5/34pyrite (iron disulphide), 5/29–30, 5/32salt contamination, 5/20–1strength and durability problems, 5/29, 8/8sulfates, 8/12–13unsound particles, 5/29–34

Aggregate properties and effects on concrete:about aggregates, 8/2density of aggregate, 8/8–9drying shrinkage, 8/9–10fines content, 8/11–12history of aggregates, 8/1–2impurities, 8/12–13large particles, 8/6particle size, 8/3–5shape of particles, 8/6–8soundness of aggregates, 8/10–11strength of aggregate, 5/29, 8/8surface texture of particles, 8/6–8thermal properties of expansion, 8/11water absorption, 8/9

Aggregate prospecting and processing seeLimestone; Sands and gravels

Air-entraining admixtures:about air-entraining admixtures, 4/21–2bedding mortars and renders, 4/25bleed reduction effects, 4/24compaction/cohesion of low-workability

mixes, 4/24–5factors affecting air entrainment, 4/22–3freeze-thaw resistance, 4/23–4see also Admixtures

Alkali-metal oxides, influence of, 1/13Alkali-silica reaction (ASR):

and fly ash, 3/13–14ggbs based concrete, 3/31–2silica fume based concrete, 3/48suppression of, metakaolin based concrete,

3/56Alkaline hydrolysis, CACs, 2/21Alluvial deposits, 5/16–17, 5/18

Alucrete, 5/21Alumina ratio (AR), 1/11–12Aluminates, hydration of, 1/17ANFO (ammonium nitrate ferrous oxide), 6/10Anhydrous phases, definition, 2/3Anti-bleed admixtures, 4/34Anti-washout/underwater admixtures:

about underwater admixtures, 4/29applications, 4/30mechanism, 4/29see also Admixtures

AR (alumina ratio), 1/11–12ASR see Alkali-silica reactionast (American Society for Testing and Materials),

1/23

Bacterial resistance, CACs, 2/23–4Ball mills, 1/14Bauxite, 5/21Binders, CEN TC 51 standards for construction

binders, 1/32Biotite as aggregate impurities, 5/34Bogue C3S to C2 ratio measurement method,

1/11Bonding, silica fume based concrete, 3/44Boulder clay, 5/18Brittleness, with silica fume, 3/44

CAA (Cement Admixtures Association), 4/34CACs (calcium aluminate cements):

about CACs, 2/1–3abrasion resistance, 2/24–5acid/sewage resistance, 2/23–4alkaline hydrolysis, 2/21applications, 2/1–2assurance of conformity, 2/18bacterial resistance, 2/23–4chemistry and mineralogy, 2/3–11

zone of composition, 2/3–4clinker microstructure, 2/5conversion, impact of, 2/9–11conversion inhibiting effects, 2/19dimensional stability, 2/28–9durability/resistance to degradation, 2/19–21ettringite systems stability, 2/28–9flash setting, 2/26–7in floor levelling compounds, 2/27–8freeze-thaw damage, 2/21heat evolution, 2/13–14hydration and conversion, 2/6–9, 2/15–16impact resistance, 2/24–5

Index I/3

insulated moulds, 2/18metastable phases, 2/7microstructures, 2/7–8, 2/10, 2/11in mixed binder systems, 2/26–8modern uses, 2/23–5phases, 2/4–6plasticizers and superplasticizers, 2/13rapid strength development, 2/25rate of reaction, 2/13–14reinforcement corrosion, 2/20setting times, 2/12shrinkage, 2/14–15shrinkage compensation with ettringite–

containing systems, 2/28–9strength development, 2/15–19structural collapse problems, 2/21–2sulphate attack, 2/20–1terminology, 2/3thermal resistance, 2/25water-to-cement ration considerations,

2/16–18workability, 2/12–13

Carbonate rocks, 5/9Carbonation, silica fume based concrete, 3/47Carbonization of concrete:

and fly ash, 3/14ggbs based concrete, 3/29–30

Cement see Cement grinding; Cementparameters; Cement quality controlmethods; Composite cements; Portlandcement; Portland cement hydration;Portland cement manufacture; Standards

Cement grinding, 1/14–15ball mills, 1/14efficiency considerations, 1/14–15open-circuit mills, 1/14

Cement parameters, effects on properties withquality control:

key parameters, 1/36setting time, 1/37strength development, 1/37–41

clinker alkalis and SO3, 1/38–9compound composition, 1/39–40fineness, 1/37free lime, 1/39loss on ignition (LOI), 1/37–8SO3 level and forms of SO3 present,

1/40–1water demand (workability), 1/36see also Cement quality control methods

Cement paste, hydrated, 1/22, 1/23Cement quality control methods, 1/33–6

Blaine cell surface area measurement, 1/35continuous on-line analysis, 1/35EN 196-1 mortar test, 1/35Vicat needle setting time test (EN 196-3),

1/35wet chemical analysis (BS EN 192-2),

1/33–4X-ray fluorescence, 1/34

Chalk particles as aggregate impurities, 5/33Chloride:

chloride ion penetration resistance:ggbs based concretes, 3/32metakaolin based concrete, 3/54–5

influence of, 1/13resistance of silica fume based concrete,

3/46–7sea water attack on reinforcing, 3/14–15

Clastic (detrital) rocks, 5/8Clinker, definition, 2/3Clinker alkalis and SO3, effect on strength,

1/38–9Clinker grinding see Cement grindingClinker manufacture:

alkali metals, influence of, 1/13alumina ratio (AR), 1/11–12basic process, 1/7Bogue content ratio measurement method,

1/11, 1/12chemical reactions, 1/8–10composition calculation, 1/12control ratios, 1/10–12crystallization, 1/10European clinkers, 1/9fluorine, influence of, 1/13lime saturation factor (LSF), 1/10–12minor constituents, 1/12–13photomicrographs of clinker, 1/10precalciner dry process kiln, 1/6–8primary fuels, 1/6primary/secondary raw materials, 1/6raw materials, 1/5–6rotary kilns, 1/6–8silica ratio (SR), 1/10–12trace metals, influence of, 1/13see also Portland cement manufacture

CLSM (controlled low-strength material)/foamedconcrete, 4/31–2

Coal particles as aggregate impurities, 5/33–4Coastal deposits of sand/gravel, 5/18Comb polymers (PCEs), 4/12Composite cements:

about composite cements, 1/25–7

I/4 Index

European usage, 1/26with finely ground limestone, 1/26with fly ash, 1/27with granulated blastfurnace slag, 1/26

Concrete:with lightweight aggregates, 7/2–3see also Aggregate properties and effects on

concreteCondensed silica fume see Silica fume with

cement/concreteConversion, CACs, 2/6–11Corrosion-inhibiting admixtures:

about corrosion inhibitors, 4/26–7applications, 4/27–8mechanism, 4/27see also Admixtures

Creep:fly ash effects, 3/11–12ggbs based concrete, 3/26metakaolin based concrete, 3/53silica fume based concrete, 3/45

Crushed rock:about crushed rock, 5/12–14UK locations, 5/15see also Aggregate classification and sources;

Rocks/rock materialCuring:

fly ash effects, 3/12–13ggbs based concrete, 3/27

DEF (delayed ettringite formation), 1/31Density of aggregates:

dry loose density, 7/1–2particle density, 7/1

Dispersing admixtures:about dispersing admixtures, 4/9chemical structure, 4/10–12electrostatic dispersion, 4/13–14flocs, 4/13mechanism of dispersion, 4/13–15normal plasticizers, 4/9–10, 4/15–16steric stabilization, 4/14–15superplasticizers, 4/10, 4/16–17see also Admixtures

Dissolution process, CACs, 2/6Dolomite, 5/8–9Dormant period, 1/16Dry loose density of aggregates, 7/1–2Drying shrinkage see ShrinkageDurability of concrete:

CAC based, 2/19–21

fly ash effects, 3/13ggbs based, 3/27–34

E modulus, 3/44Earth’s crust, 5/4EFCA (European Federation of Concrete

Admixture Associations), 4/34Efflorescence:

metakaolin based concrete, 3/57silica fume based concrete, 3/47

Elastic modulus, effects of fly ash, 3/11EN standards see StandardsErosion and abrasion, silica fume based concrete,

3/45Ethylene glycol derivatives, 4/28Ettringite formation, 1/17

and admixtures, 4/8with CACs, 2/28–9

European Standard for Common cements:about EU standards, 1/22–3see also Standards

Expanded clay aggregate, 7/8–9

Ferrocrete, 5/21Fire resistance, silica fume based concrete, 3/45Flash setting, with CACs, 2/26–8Flocs, in dispersing admixtures, 4/13Floor levelling compounds, 2/27Fluorine, influence of, 1/13Fluvio-glacial deposits of sand/gravel, 5/18Fly ash:

about fly ash, 3/3–4and the alkali silica reaction (ASR), 3/13–14and carbonization of concrete, 3/14in composite cement, 1/25–7concrete durability effects, 3/13creep effects, 3/11–12curing effects, 3/12–13elastic modulus effects, 3/11fineness, 3/8and freeze-thaw damage, 3/15and heat of hydration, 3/9–10high fly ash content concrete, 3/17particle shape and density, 3/6–7, 3/8and pozzolanic reactions, 3/4–6and resistance to acids, 3/17roller-compacted concrete with high fly ash

content, 3/17and sea water/chloride attack on reinforcing,

3/14–15setting time effects, 3/11sprayed concrete with pfa, 3/17–18

Index I/5

strength considerations for concrete, 3/8–9and sulphate attack, 3/16tensile strain effects, 3/12and thaumasite form of sulfate attack, 3/16–

17thermal expansion coefficient effects, 3/12variability problems, 3/7

Foamed concrete/CLSM admixtures, 4/31–2Foamed slag aggregate, 7/6Free lime, effect on parameters, 1/39Freeze-thaw damage/resistance:

aggregate deleterious material problems,5/29

air-entraining admixtures, 4/23–4CACs, 2/21and fly ash, 3/15ggbs based concrete, 3/34metakaolin based concrete, 3/54silica fume based concrete, 3/47–8

Geology:alluvial deposits, 5/16–17Earth’s crust, 5/4–5Geological Column, 5/4glacial and preglacial regions, 5/17–19hot desert regions, 5/20–1river terrace deposits, 5/18temperate fluvial environments, 5/16–17tropical hot wet environments, 5/21, 5/22UK sources of aggregates, 5/15see also Rocks/rock material

ggbs (ground granulated blastfurnace slag)based concretes:

abrasion resistance, 3/34alkali-silica reaction (ASR), 3/31–2alkalinity, 3/34blended cement production, 3/21–2carbonation, 3/29–30chloride ingress, 3/32–3curing, 3/27durability, 3/27–34formwork pressures and striking times, 3/27freeze-thaw resistance, 3/34hardened concrete:

compressive strength and strengthdevelopment, 3/23–5

creep, 3/26–7drying shrinkage, 3/27elastic modulus, 3/25–6temperature influence, 3/25tensile strength, 3/25

permeability and porosity, 3/28–9plastic concrete:

bleeding and settlement, 3/22–3heat of hydration and early age thermal

cracking, 3/23water demand/workability, 3/22

with silica fume, 3/42–3sulfate resistance, 3/32surface finish, 3/27

ggbs (ground granulated blastfurnace slag)material/properties, 3/18–34

about ggbs, 3/18chemical composition, 3/19, 3/20–1chemical reactions, 3/19–21fineness, 3/21glass count, 3/21granulation/pelletization, 3/19hydration mechanism, 3/20production, 3/18–19reaction rate, 3/20

Glacial and preglacial regions/deposits, 5/17–19, 5/18

Granite, weathering grades, 5/12, 5/13Granulated blastfurnace slag, in composite

cement, 1/25–7Granulation and pelletization of ggbs, 3/19Gravels see Sands and gravelsGround granulated blastfurnace slag (ggbs) see

ggbsGrout admixtures, 4/34

Hardened concrete:ggbs cement based, 3/23–7with silica fume, 3/43–8

Haydite expanded aggregate, 7/3Health and safety:

admixtures, 4/36cement, 1/43–5silica fume, 3/35–7

Hot desert regions, aggregates, 5/20–1Hydrates/hydration:

and conversion, CACs, 2/6–9definition, 2/3fly ash effects, 3/9–10ggbs mechanism, 3/20hydrated cement paste, 1/22, 1/23see also Portland cement hydration

Igneus rocks, 5/6, 5/7Impact resistance, CACs, 2/24–5IR (insoluble residue) for cements, 1/31

I/6 Index

Kaolin, 3/49see also Metakaolin

Kilns, 1/6–8

LECA (lightweight expanded clay aggregate),7/3

Liapor, 7/4Lightweight aggregate:

about lightweight aggregates, 7/1–2definition, 7/1–2expanded clay aggregate, 7/8–9

future of, 7/11expanded shale aggregate, 7/9–10expanded slate aggregate, 7/10foamed slag aggregate, 7/6Haydite expanded aggregate, 7/3history of production, 7/3–4investment considerations, 7/4LECA (lightweight expanded clay aggregate),

7/3Liapor, 7/4Lytag, 7/4particle and dry loose density, 7/1–2pelletized expanded blastfurnace slag

aggregate, 7/6–7pelletized pulverized–fuel ash (PFA), 7/3–4,

7/5future of, 7/10–11production methods and techniques, 7/5–10resource materials, 7/4–5sintered pulverized fuel ash (PFA), 7/7–8

future of, 7/11slag pelletization, 7/6–7in structural concrete, 7/2–3

Lignin and sulphonated lignin, 4/10–11Lignite particles as aggregate impurities,

5/33–4Limestone, 3/58–9, 5/8–9

allochemical constituent particles, 5/9extraction and processing with explosives,

6/10–11Folk classification, 5/10lime saturation factor (LSF), 1/10–12limestone filler, 3/58see also Aggregate classification and sources

LOI (loss on ignition) for cements, 1/31effect on parameters, 1/37–8

LSF (lime saturation factor), 1/10–12Lytag, 7/4

Marine deposits of sand/gravel, 5/18

Metakaolin:Agrément Certificate Number, 3/51compatibility with blended cements, 3/57efflorescence, 3/57occurrence and extraction, 3/49physical properties, 3/49production, 3/49properties of Metakaolin based cement/

concrete:acid resistance, 3/55alkali-silica reaction suppression, 3/56chloride ion penetration resistance,

3/54–5compressive strength, 3/51–2creep strain, 3/53drying shrinkage, 3/53durability, 3/54–6freeze/thaw resistance, 3/54fresh concrete, 3/51hardened concrete, 3/51–3sulfate resistance, 3/54tensile strength, 3/52–3

reaction mechanisms, 3/49–51Metamorphic rocks, 5/6Metastable phases, CACs, 2/7Micas as aggregate impurities, 5/34Microsilica see Silica fume with cement/concreteMicrostructures, CACs, 2/5, 2/7–8, 2/10, 2/11Minerals:

definition, 5/5see also Rocks/rock material

Mixed binder systems, with CACs, 2/26–8Mortar admixtures, 4/33Mortar tests (EN 196–1):

and field concrete results, 1/42and quality control, 1/35

Muscovite as aggregate impurities, 5/34

Nucleation process, CACs, 2/6–7

Open-circuit mills, 1/14

Particle density of aggregates, 7/1PCEs (polycarboxylate ethers), 4/12, 4/16Pelletization and granulation of ggbs, 3/19Pelletized expanded blastfurnace slag aggregate,

7/6–7Permeability:

ggbs based concrete, 3/28–9silica fume based concrete, 3/45–6

Index I/7

PFA (pelletized pulverized-fuel ash), 3/17–18,7/3–4, 7/5

sintered PFA, 7/7–8Phase, definition in cement applications, 2/3Plastic concrete, ggbs cement based, 3/22–3Plasticizers:

with CACs, 2/13normal plasticizers, 4/9–10, 4/15–16retarding plasticizers, 4/17–20with silica fume, 3/40see also Dispersing admixtures;

SuperplasticizersPNS (polynaphthalene sulphonate), 4/11Polycarboxylate ethers (PCEs), 4/12, 4/16Polymer dispersion admixtures, 4/32Pore blocking, 4/25, 4/26Portland cement:

applications for various types, 1/42–4composite cements, 1/25–7delayed ettringite formation (DEF), 1/31European type usage, 1/30health and safety considerations, 1/43–5insoluble residue (IR), 1/31loss on ignition (LOI), 1/31main types summary, 1/24‘pure’ types, 1/24–5and sulphate-resisting concrete, 1/25see also Cement parameters, effects on

properties with quality control; Cementquality control methods; Portlandcement hydration; Portland cementmanufacture; Standards

Portland cement hydration, 1/15–22about hydration, 1/15dehydrated gypsum, effects of, 1/20–1dormant period, 1/16ettringite formation, 1/17heat release/temperature rise considerations,

1/19hydrated cement paste, 1/22, 1/23hydration of aluminates, 1/17hydration of C3A and C4AF, 1/16–17hydration process, 1/17–20hydration of silicates, 1/15–16optimization of level of rapidly soluble

calcium sulfate, 1/20–1outer hydration product, 1/18particle size, 1/19–20setting time, 1/17–18techniques for study of, 1/21–2

Portland cement manufacture:chemical content, 1/8–10

terminology (C, S, A and F), 1/5

history of manufacture, 1/3–5clinker control, 1/4–5early calcium silicate cements, 1/4Joseph Aspdin patent, 1/4production landmarks, 1/5rotary kiln introduction, 1/4

minor constituents, influence of, 1/12–13raw materials, 1/6white cement clinker, 1/25see also Cement grinding; Clinker

manufacture; Portland cement hydrationPozzolanic materials:

natural pozzolanas, 3/18pozzolanic fly ash cement, 1/31pozzolanic reaction and concrete, 3/3–6, 3/8

Pre-cast, semi-dry admixtures, 4/32Precalciner dry process kiln, 1/6–8Precipitation process, CACs, 2/6Pumping aids, 4/30Pyrite (iron disulphide) in aggregates, 5/29–30

Quality control:aggregates, quarry assessment, 5/23–5cement see Cement quality control methods

Quarry assessment of aggregates, 5/23–5

Ready-to-Use Mortar, 4/33Reinforcement:

attack by sea water/chloride, 3/14–15corrosion of, and CACs, 2/20

Retarding admixtures:about retarding admixtures, 4/17applications, 4/19–20mechanism of retardation, 4/12–13performance, 4/19–20set retardation, 4/19workability retention, 4/18see also Admixtures

Rheology/rheology measurements, 4/6–7River terrace deposits, 5/18Rocks/rock material:

about rock, 5/3–6carbonate rocks, 5/9clastic (detrital) rocks, 5/8dolomite, 5/8–9geological classification, 5/6–12granite, 5/12, 5/13igneus rocks, 5/6, 5/7limestones, 5/8–9

allochemical constituent particles, 5/9Folk classification, 5/10

I/8 Index

metamorphic rocks, 5/6–8sedimentary rocks, 5/6, 5/7, 5/8–9UK locations, 5/15weathering grades, 5/9, 5/11, 5/12, 5/13

Roller-compacted concrete with high fly ashcontent, 3/17

Ronan Point collapse with CAC concrete, 2/22Rotary kilns, 1/6–8

Salt contamination, aggregates, 5/20–1Sands and gravels, extraction:

about extraction, 6/2dredging at anchor, 6/4dry ‘pit’ extraction, 6/3–4marine extraction, 6/4trail dredging, 6/4–5wet extraction, 6/2–3see also Aggregate classification and sources

Sands and gravels, features and geology:advantages, 5/14–16alluvial deposits, 5/16, 5/18boulder clay, 5/18coastal deposits, 5/18fluvio-glacial deposits, 5/18glacial deposits, 5/18marine deposits, 5/18salt contamination, 5/20–1sources, 5/12UK locations, 5/15

Sands and gravels, processing:blinding, 6/9classification, 6/9crushing:

about crushing, 6/5–6gyratory (cone) crushers, 6/7impact crushers, 6/7–8jaw crushers, 6/6–7

de-watering, 6/9–10pegging, 6/9screening, 6/8–9washing, 6/9

Sea water/chloride attack on reinforcing, 3/14–15

Sedimentary rocks, 5/6, 5/7, 5/8–9SEM (scanning electron microscopy) study

technique, 1/22Setting time:

CACs, 2/12and clinker reactivity and cement fineness,

effects of, 1/37fly ash effects, 3/11

Sewage/acid resistance, CACs, 2/23–4Shrinkage:

CACs, 2/14–15drying shrinkage of aggregates, 8/9–10ggbs based concrete, 3/27metakaolin based concrete, 3/53silica fume based concrete, 3/44–5

Shrinkage-reducing admixtures:about shrinkage reduction, 4/28applications, 4/29mechanism, 4/28–9see also Admixtures

Silica fume with cement/concrete:about silica fume, 3/34bleedwater effects, 3/40blended cement mixes, 3/42–3characteristics, 3/35, 3/37densified form, 3/37effects on fresh concrete, 3/39–40effects on setting and hardening, 3/41–2fire resistance, 3/45ggbs blends, 3/42–3hardened concrete properties, 3/43–8

abrasion and erosion, 3/45alkali-silica reaction, 3/48bonding, 3/44brittleness and E modulus, 3/44carbonation, 3/47chloride resistance, 3/46–7compressive strength, 3/43creep, 3/45efflorescence, 3/47frost resistance, 3/47–8shrinkage, 3/44–5sulfate resistance, 3/46tensile and flexural strength, 3/43

health and safety, 3/35–7micropelletized, 3/38mixed design criteria, 3/40–1national standards and specifications, 3/38–

9permeability, 3/45–6with plasticizers and superplasticizers, 3/40production and extraction, 3/35–6in readymix production, 3/41as slurry, 3/38with slurry material, 3/40undensified form, 3/37

Silica ratio (SR), 1/10–12Sintered pulverized fuel ash (sintered PFA),

7/7–8Slag pelletization, 7/6–7

Index I/9

Sprayed concrete:admixtures, 4/31with pfa and fly ash, 3/17–18

SR (silica ratio), 1/10–12Standards:

about EU standards, 1/23, 1/27–32ASTM-based standards, 1/32–3CEN TC 51 standards for construction binders,

1/32EN 196 (test methods), 1/23EN 196-1 (compressive strength), 1/29–30,

1/35EN 196-2 (wet chemical analysis methods),

1/33–4EN 196-3 (Vicat needle test), 1/35EN 196-5 (pozzolanicity of fly ash), 3/8EN 197-1 (common cements), 1/23, 1/27–32

chemical requirements, 1/31compositional requirements, 1/29strength setting time and soundness

requirements, 1/30testing frequencies, 1/32types and compositions permitted, 1/28

EN 197-2 (EU, assessment for conformity),1/23

EN 206-1 (additions, lightweight concrete),3/58–9, 7/1

EN 932-1 and 3 (aggregates), 5/23, 8/3EN 934-2 (plasticizing/accelerating/

hardening), 4/4EN 934-3 (retarded ready-to-use mortars),

4/4EN 934-4 (grout for prestressing), 4/4EN 934-5 (sprayed concrete), 4/4EN 12620 (aggregates), 5/14, 8/12EN 13055 (lightweight aggregate), 7/1ENV 197-1 (slag elements), 3/22Portland cement, 1/22–3silica fume national standards, 3/38–9

Steric stabilization, 4/14–15Strength:

aggregate deleterious material problems,5/29

CACs, 2/15–19and cement fineness, 1/37and clinker alkalis and SO3, 1/38–9and compound composition, 1/39–40compressive:

with metakaolin, 3/51–2with silica fume, 3/43, 3/44

and free lime, 1/39

and LOI (loss on ignition), 1/37–8and quality control, 1/37–41rapid strength development CACs, 2/25and SO3 levels, 1/40–1tensile and flexural:

with metakaolin, 3/52–3with silica fume, 3/43

Structural collapses with CAC concrete, 2/21–2Sulfate attack/resistance:

CACs, 2/20–1fly ash effects, 3/16ggbs based concretes, 3/32metakaolin based concrete, 3/54Portland cement based concrete, 1/25silica fume based concrete, 3/46thaumasite form, 3/16–17

Sulphonated melamine formaldehydecondensates (SMF), 4/12, 4/16

Sulphonated naphthalene formaldehydecondensate (SNF), 4/11, 4/16

Superplasticizers, 4/10, 4/16–17with CACs, 2/13retarding superplasticizers, 4/17–20see also Dispersing admixtures

Tensile strain capacity, fly ash effects, 3/12Terminology, cement, 2/3Thaumasite form of sulfate attack, and fly ash,

3/16–17Thermal analysis study technique, 1/22Thermal expansion coefficient, fly ash effects,

3/12Thermal resistance, CACs, 2/25Tora de la Peira collapse with CAC concrete,

2/22Trace metals, influence of, 1/13Tropical hot wet environments, 5/21, 5/22Truck washwater admixtures, 4/32–3

Underwater/anti-washout admixtures:about underwater admixtures, 4/29applications, 4/30mechanism, 4/29see also Admixtures

Vicat needle test (EN 196-3), 1/35Volatized silica see Silica fume with cement/

concrete

I/10 Index

Water demand (workability), 1/36Water resisting (waterproofing) admixtures:

about waterproofing, 4/25application, 4/26mechanism, 4/26pore blocking, 4/25, 4/26selection, 4/26see also Admixtures

Water-reducing admixtures, 4/8see also Admixtures

Wet chemical analysis of cement(BS EN 192-2), 1/33–4

White cement clinker, 1/25Workability, CACs, 2/12–13

X-ray diffraction:glass count measurement, 3/21study technique, 1/22

X-ray fluorescence cement analysis, 1/34

Zeta potential measurements, 4/13–14